Initial add of defaria.com
[clearscm.git] / defaria.com / blogs / Status / archives / 2005_03.html
1 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
2    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
3 <html xmlns="http://www.w3.org/1999/xhtml" id="sixapart-standard">
4 <head>
5    <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
6    <meta name="generator" content="Movable Type 5.2.3" />
7
8    <link rel="stylesheet" href="http://defaria.com/blogs/Status/styles-site.css" type="text/css" />
9    <link rel="alternate" type="application/atom+xml" title="Atom" href="http://defaria.com/blogs/Status/atom.xml" />
10    <link rel="alternate" type="application/rss+xml" title="RSS 2.0" href="http://defaria.com/blogs/Status/index.xml"$>" />
11
12    <title>Status for Andrew DeFaria: March 2005 Archives</title>
13
14    <link rel="start" href="http://defaria.com/blogs/Status/" title="Home" />
15    <link rel="prev" href="http://defaria.com/blogs/Status/archives/2005_02.html" title="February 2005" />
16    <link rel="next" href="http://defaria.com/blogs/Status/archives/2005_04.html" title="April 2005" />
17 </head>
18 <body class="layout-one-column">
19    <div id="container">
20       <div id="container-inner" class="pkg">
21
22          <div id="banner">
23             <div id="banner-inner" class="pkg">
24                <h1 id="banner-header"><a href="http://defaria.com/blogs/Status/" accesskey="1">Status for Andrew DeFaria</a></h1>
25                <h2 id="banner-description">Searchable status reports and work log</h2>
26             </div>
27          </div>
28
29          <div id="pagebody">
30             <div id="pagebody-inner" class="pkg">
31                <div id="alpha">
32                   <div id="alpha-inner" class="pkg">
33                      
34                      <p class="content-nav">
35                         <a href="http://defaria.com/blogs/Status/archives/2005_02.html">&laquo; February 2005</a> |
36                         <a href="http://defaria.com/blogs/Status/">Main</a>
37                         | <a href="http://defaria.com/blogs/Status/archives/2005_04.html">April 2005 &raquo;</a>
38                      </p>
39                      
40                      
41                      
42
43                      <h2 class="date-header">March 31, 2005</h2>
44                      <a id="a000327"></a>
45                      <div class="entry" id="entry-327">
46                         <h3 class="entry-header">Finalized HybridOS</h3>
47                         <div class="entry-content">
48                            <div class="entry-body">
49                               <ul>
50   <li>Finished up CR 1 for HybridOS checkin - assigned to Thu for review</li>
51 </ul>
52                               
53                               <p class="entry-footer">
54                                  <span class="post-footers">Posted by  at  4:10 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000327.html">Permalink</a>
55                                  
56                                  
57                               </p>
58                            </div>
59                         </div>
60                      </div>
61                      
62                      
63
64                      <h2 class="date-header">March 30, 2005</h2>
65                      <a id="a000326"></a>
66                      <div class="entry" id="entry-326">
67                         <h3 class="entry-header">HybridOS Binary Comparison problems</h3>
68                         <div class="entry-content">
69                            <div class="entry-body">
70                               <ul>
71   <li>Managed to perform binary comparison of HybridOS</li>
72 </ul>
73                               
74                               <h3>HybridOS Binary Comparison</h3>
75
76 <p>HybridOS has been checked in and built. As you know the binary comparison procedure described in the GS: LOS178 Impact Summary discovered more differences. Specifically 227 .o files had differences. Further investigation revealed that the action of committing the sources to CVS caused $Header/ident strings to change. The following describes the changes to the $Header strings due to cvs commit:</p>
77
78 <div class="code"><pre>
79 tomcat:strings -a orig.uipc_usrreq.o | grep Header
80 $Header: /cvs/<font color="blue"><b>los178-cvs</b></font>/los178/sys/networking/tcpip/general/uipc_usrreq.c,v  <font color="orange"><b>1.1.1.1</b></font> <font color="green"><b>2004/03/03 00:59:24</b></font> <font color="purple"><b>emooring</b></font> Exp $
81 tomcat:strings -a new.uipc_usrreq.o | grep Header
82 $Header: /cvs/<font color="blue"><b>hybrid-os-cvs</b></font>/los178/sys/networking/tcpip/general/uipc_usrreq.c,v <font color="orange"><b>1.1</b></font> <font color="green"><b>2005/03/30 00:39:03</b></font> <font color="purple"><b>adefaria</b></font> Exp $
83 </pre></div>
84
85 <p>The changes are as follows:</p>
86
87 <ol>
88   <li>CVS Repository name changed from los178-cvs -> hybrid-os-cvs (<font color="blue">blue</font>)</li>
89
90   <li>Revision changed from whatever it was -> 1.1 (<font color="orange">orange</font>) All revisions for HybridOS are now 1.1</li>
91
92   <li>Date changed to reflect time of cvs commit to new CVS repository (<font color="green">green</font>)</li>
93
94   <li>User changed from whatever it was -> adefaria (<font color="purple">purple</font>) since I was the user to perform the commit</li>
95 </ol>
96
97 <p>Using objdump once again to disassemble these .o files and comparing the output left us with the following .o files that were still different:</p>
98
99 <ol>
100   <li>/sys/lib/libcsp_970.a (context_asm.o)</li>
101
102   <li>/sys/lib/libcsp_970.a (csp_cpu_asm.o)</li>
103
104   <li>/sys/lib/libcsp_970.a (flih.o)</li>
105
106   <li>/sys/lib/libcsp_970.a (fpu_asm.o)</li>
107
108   <li>/sys/lib/libcsp_970.a (launch_asm.o)</li>
109
110   <li>/sys/lib/libcsp_970.a (tlbmiss.o)</li>
111 </ol>
112
113 <p>Closer examination of these .o files reveals that the also contained ident strings in the text segment that had the same differences as the $Header differences described above. In other words the code was the same but the version strings and dates changed, as is expected.</p>
114
115                               
116                               <p class="entry-footer">
117                                  <span class="post-footers">Posted by  at  4:59 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000326.html">Permalink</a>
118                                  
119                                  
120                               </p>
121                            </div>
122                         </div>
123                      </div>
124                      
125                      
126
127                      <h2 class="date-header">March 29, 2005</h2>
128                      <a id="a000325"></a>
129                      <div class="entry" id="entry-325">
130                         <h3 class="entry-header">Hybrid OS</h3>
131                         <div class="entry-content">
132                            <div class="entry-body">
133                               <ul>
134   <li>Checked HybridOS into CVS</li>
135
136   <li>Rebuilt HybridOS</li>
137
138   <li>Attempted binary comparison - fails due to $Header$ strings</li>
139 </ul>
140                               
141                               <h3>Binary Differences</h3>
142
143 <p>Well the build finished but the binary comparison as per the Impact Summary failed. For a while I thought I did something wrong so I went back and re-extracted from the SCL and rebuild the old LOS178 that I had stored on the side, etc. Still it kept failing! Not only 25 files that were different and needed to be disassembled and compared but more like 228 files! What's going on?!?</p>
144
145 <p>So I dug deeper... Seems that $Header is embedded in some .o files and the $Headers differ (picking at random a .o that didn't compare):</p>
146
147 <div class="code"><pre>
148 tomcat:strings -a new.uipc_usrreq.o | grep Header
149 $Header: /cvs/hybrid-os-cvs/los178/sys/networking/tcpip/general/uipc_usrreq.c,v 1.1 2005/03/30 00:39:03 adefaria Exp $
150 tomcat:strings -a orig.uipc_usrreq.o | grep Header
151 $Header: /cvs/los178-cvs/los178/sys/networking/tcpip/general/uipc_usrreq.c,v 1.1.1.1 2004/03/03 00:59:24 emooring Exp $
152 tomcat:
153 </pre></div>
154
155 <p>So as you can see, we have differences. I don't know why all 2543 .o files extracted from the .a files didn't all differ.</p>
156                               
157                               <p class="entry-footer">
158                                  <span class="post-footers">Posted by  at 11:08 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000325.html">Permalink</a>
159                                  
160                                  
161                               </p>
162                            </div>
163                         </div>
164                      </div>
165                      
166                      
167
168                      <h2 class="date-header">March 28, 2005</h2>
169                      <a id="a000324"></a>
170                      <div class="entry" id="entry-324">
171                         <h3 class="entry-header">HybridOS built</h3>
172                         <div class="entry-content">
173                            <div class="entry-body">
174                               <ul>
175   <li>Built Hybrid OS for GD and performed binary comparison</li>
176
177   <li>Adding files to CVS on Tomcat</li>
178
179   <li>Completed GD LOS178 Impact Summary</li>
180
181   <li>Resolved long standing issue regarding gnuaout vs. gnu</li>
182 </ul>
183                               
184                               <h2>GD LOS178 Impact Summary</h2>
185
186 <h3>Export sources from LOS178</h3>
187
188 <p>Export sources from LOS178 CVS tree using the CVS tag REL_LOS178_2p0p0_ppc_FCS. The export will come from the machine named Rock using CVSROOT=:pserver:anoncvs@rock:/cvs/los178-cvs:</p>
189
190 <div class="code"><pre>
191 tomcat:export CVSROOT=:pserver:anoncvs@rock:/cvs/los178-cvs
192 tomcat:cvs login
193 Logging in to :pserver:anoncvs@rock:2401/cvs/los178-cvs
194 CVS password:
195 tomcat:cvs export -r REL_LOS178_2p0p0_ppc_FCS los178
196 </pre></div>
197
198 <h3>Extract prebuilt CDK</h3>
199
200 <p>Extract prebuilt CDK (sunos-xcoff-ppc) binary also using the tag of REL_LOS178_2p0p0_ppc_FCS from Rock. Note that this prebuilt CDK comes from the bin-image section of the CVS repository and that we are only using the ppc.cdksol.tar.gz image:</p>
201
202 <div class="code"><pre>
203 tomcat:cvs export -r REL_LOS178_2p0p0_ppc_FCS bin-image/ppc.cdksol.tar.gz
204 </pre></div>
205
206 <h3>Extract other tools</h3>
207
208 <p>The package.sh script from the toolbox area is used to package up the images so we need to extract that too:</p>
209
210 <div class="code"><pre>
211 tomcat:cvs export -r REL_LOS178_2p0p0_ppc_FCS toolbox/package.sh
212 </pre></div>
213
214 <h3>Test Build</h3>
215
216 <p>Perform test build.</p>
217
218 <div class="note">
219 Note: Test build requires a symlink from /usr/lynx/3.1.0/ppc/cdk/sunos-xcoff-ppc/bin/bison.simple -> $ENV_PREFIX/cdk/sunos-xcoff-ppc-/bin/bison-simple due to a hard coded path dependency.  This path was not modified due to LOS-178 RSC restrictions
220 </div>
221
222 <p>Steps performed are:</p>
223
224 <ol>
225   <li>Create ppc_dev area to perform the build in:<br>
226
227 <div class="code"><pre>
228 tomcat:mkdir ppc_dev
229 </pre></div></li>
230
231   <li>Copy in sources:<br>
232
233 <div class="code"><pre>
234 tomcat:rsync -a los178 ppc_dev
235 </pre></div></li>
236
237   <li>Unpack CDK into build area<br>
238
239 <div class="code"><pre>
240 tomcat:cd ppc_dev
241 tomcat:gnutar -zxpf ../bin-image/ppc.cdksol.tar.gz
242 </pre></div></li>
243
244   <li>Perform build:<br>
245
246 <div class="code"><pre>
247 tomcat:make DEVELOPMENT=yes install > install.log
248 </pre></div></li>
249
250   <li>Check install.log for errors</li>
251 </ol>
252
253 <h3>Perform binary comparison test</h3>
254
255 <p>This binary comparison test is different from the normal binary comparison tests. Basically we are simply extracting all .o's from all .a's in the packaged versions of the product. A little utility script was written to find all .a libraries and copy them to an area (complibs) broken out by the path to the library, then extract all .o's from the .a's. This script is called unpack_libs. It is not intended that such a comparison be performed on a regular basis so this script is more of a one shot script.</p>
256
257 <p>Further, a build will create a lot of libraries but not all libraries created will be packaged and shipped. Since we are comparing against a previously built and packaged release we must package up and unpack the build we just performed. This is done using the toolbox/package.sh script as follows:</p>
258
259 <ol>
260   <li>Package up the image just built:<br>
261
262 <div class="code"><pre>
263 tomcat:toolbox/package.sh ppc_dev dev
264 </pre></div></li>
265
266     <li>Unpack images to new area:<br>
267
268 <div class="code"><pre>
269 tomcat:mkdir new
270 tomcat:cd new
271 tomcat:for tarfile in ../media/*.tar.gz; do
272 > gnutar -zxpf $tarfile
273 > done
274 </pre></div></li>
275
276   <li>Gather all libraries and extract their .o's:<br>
277
278 <div class="code"><pre>
279 tomcat:mkdir complibs
280 tomcat:../unpack_libs
281 </pre></div></li>
282
283   <li>Unpack old images (from  t3:/export/scl/los178/2p0p0/FCS/) to old area:<br>
284
285 <div class="code"><pre>
286 tomcat:cd ..
287 tomcat:mkdir old
288 tomcat:cd old
289 tomcat:# copy old tar images here
290 tomcat:for tarfile in *.tar.gz; do
291 > gnutar -zxpf $tarfile
292 > done
293 </pre></div></li>
294
295   <li>Gather all libraries and extract their .o's:<br>
296
297 <div class="code"><pre>
298 tomcat:mkdir complibs
299 tomcat:../unpack_libs
300 </pre></div></li>
301
302   <li>Perform diff<br>
303
304 <div class="code"><pre>
305 tomcat:cd ..
306 tomcat:diff -r old/complibs new/complibs
307 </pre></div></li>
308
309   <li>The above will result in 25 .o files being different. Use objdump -D to disassemble these files and compare the disassembled output. No differences detected in disassembled output.</li>
310 </ol>
311
312 <h3>Import sources into new CVS repository</h3>
313
314 <p>Sources  will be  imported into the CVS repository using the following command:</p>
315
316 <div class="code"><pre>
317 tomcat:cd los178
318 tomcat:export CVSROOT=:pserver:adefaria@tomcat:/cvs/hybrid-os-cvs
319 tomcat:cvs login
320 Logging in to :pserver:adefaria@tomcat:2401/cvs/hybrid-os-cvs
321 CVS password:
322 tomcat:# First add all directories
323 tomcat:find . ! -name CVS -type d -exec cvs add -m "HybridOS import from LOS178" {} \;
324 tomcat:# Now add all files
325 tomcat:find . -type f -exec cvs add -m "HybridOS import from LOS178" {} \;
326 tomcat:cvs commit
327 </pre></div>
328
329 <p>Additionally the binary CDK image was checked into binary-image:</p>
330
331 <div class="code"><pre>
332 tomcat:cd ../bin-image
333 tomcat:cvs add -m "HybridOS import from LOS178" ppc.cdksol.tar.gz
334 tomcat:cvs commit
335 </pre></div>
336
337 <p>Finally the toolbox/package.sh script as checked into toolbox:</p>
338
339 <div class="code"><pre>
340 tomcat:cd ../toolbox
341 tomcat:cvs add -m "HybridOS import from LOS178" package.sh
342 tomcat:cvs commit
343 </pre></div>
344
345 <h3>Tag initial sources with the tag REL_HYBRIDOS_1p0_ppc_20050328</h3>
346
347 <p>All sources, bin packages and toolbox scripts are then tagged:</p>
348
349 <div class="code"><pre>
350 tomcat:cvs tag REL_HYBRIDOS_1p0_ppc_20050328 los178 bin-image toolbox
351 </pre></div>
352
353 <h3>Check out all sources and prebuilt CDK and perform build procedure again</h3>
354
355 <p>Next we check out all sources, bin-image and toolbox scripts into new fresh areas and then perform the build procedure as described above.</p>
356
357 <h3>Following successful build perform binary comparison test again</h3>
358
359 <p>Perform the binary comparison described above again.</p>
360
361 <h3>Package build to archive area</h3>
362
363 <p>Use the package script to package up the images and place in the archive area at tomcat:/export/dev_archive/hybridos/1p0/20050328/solaris/media/ppc</p>
364
365 <h2>Long standing issue regarding gnuaout vs. gnu</h2>
366
367 <p>This has been bugging me for a while and I finally tracked it down. Often I'd build a toolchain then attempt to build LynxOS and it would fail when attempting to get the compiler. It seems that the toolchain build was packing up the compiler tar image with one name and the build scripts were using another name to try to find it. This resulted in errors. Now I had gotten around this via a symlink but I've been wanting to make the two build procedures agree on the names of things...</p>
368
369 <p>As Adam writes here the preferred name for the toolchain tar image is derived from config.guess:</p>
370
371 <blockquote type=cite>
372 <p>Andrew DeFaria writes:</p>
373
374   <blockquote type=cite>
375   <p>toolchain-i686-pc-linux-gnu-i386.tar.gz</p>
376   </blockquote>
377
378 <p>This. But I think we get this from config.guess so try to see how this nice level of abstraction fails before you hard-code something.</p>
379 </blockquote>
380
381 <p>The "toolchain-" portion is standard for the toolchain. The "i686-pc-linux-gnu" portion comes out of config.guess:</p>
382
383 <div class="code"><pre>
384 [int@dopey 20050207]$ /export/build1/LYNXOS_500/work_area/toolchain/3.2.2/toolchain/src/config.guess
385     i686-pc-linux-gnu
386 </pre></div>
387
388 <p>However the int_tools uses the following code to determine the name of the toolchain tar image:</p>
389
390 <div class="code"><pre>
391 proc Unload_com { platform dir comp_release format host } {
392
393   switch "$host" {
394     <font color="red"><b>"linux"  { set host_platform "i686-pc-linux-<u>gnuaout</u>" }</b></font>
395     "win32"  { set host_platform "i686-pc-cygwin" }
396     "sunos"  { set host_platform "sparc-sun-solaris2.7" }
397     "lynxos" { if { "$platform" == "x86" } {
398                    set host_platform "i386-lynx-lynxos"
399                }
400                if { "$platform" == "ppc" } {
401                      set host_platform "powerpc-lynx-lynxos"
402                }
403              }
404   }
405   if { "$platform" == "x86" } {
406         set target_platform "i386"
407   } else {
408         set target_platform "$platform"
409   }
410   set COMPILER_TAR_GZ "toolchain-$host_platform-$target_platform.tar.gz"
411 </pre></div>
412
413 <p>The highlighted portion above is the line in error and the underlined portion should change to simply "gnu". The int_tools do not have the benefit of being able to call config.guess so this could likely break in the future again.</p>
414
415 <p>I will perform this change, along with other int_tool changes required for the new tag labeling under and ECR.</p>
416
417                               
418                               <p class="entry-footer">
419                                  <span class="post-footers">Posted by  at  6:30 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000324.html">Permalink</a>
420                                  
421                                  
422                               </p>
423                            </div>
424                         </div>
425                      </div>
426                      
427                      
428
429                      <h2 class="date-header">March 25, 2005</h2>
430                      <a id="a000323"></a>
431                      <div class="entry" id="entry-323">
432                         <h3 class="entry-header">LOS178 compares</h3>
433                         <div class="entry-content">
434                            <div class="entry-body">
435                               <ul>
436   <li>LOS178 does finally compare. Apparently .o files produced by assembly must be compared at the source level by using objdump to dump assembly</li>
437
438   <li>Attempting build of LOS178 on Tomcat before putting GD LOS178 into CVS.</li>
439
440   <li>Hit problem with hard coded paths - need Jeff to resolve these</li>
441 </ul>
442                               
443                               <h3>Comparing Assembled .o files</h3>
444
445 <p>25 files did not compare. Turns out these were probably generated by the assembler. Instead we use objdump which comes in in the CDK. This eliminates differences that may be due to date/timestamps.</p>
446                               
447                               <p class="entry-footer">
448                                  <span class="post-footers">Posted by  at  9:37 AM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000323.html">Permalink</a>
449                                  
450                                  
451                               </p>
452                            </div>
453                         </div>
454                      </div>
455                      
456                      
457
458                      <h2 class="date-header">March 23, 2005</h2>
459                      <a id="a000321"></a>
460                      <div class="entry" id="entry-321">
461                         <h3 class="entry-header">LOS178 build</h3>
462                         <div class="entry-content">
463                            <div class="entry-body">
464                               <ul>
465   <li>Building LOS178 Development version to compare to previously release images</li>
466 </ul>
467                               
468                               <p class="entry-footer">
469                                  <span class="post-footers">Posted by  at  4:34 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000321.html">Permalink</a>
470                                  
471                                  
472                               </p>
473                            </div>
474                         </div>
475                      </div>
476                      
477                      
478
479                      <h2 class="date-header">March 22, 2005</h2>
480                      <a id="a000322"></a>
481                      <div class="entry" id="entry-322">
482                         <h3 class="entry-header">Bluecat build still failing</h3>
483                         <div class="entry-content">
484                            <div class="entry-body">
485                               <ul>
486   <li>Bluecat build still fails with same problem. Emailed Sasha</li>
487
488   <li>Native PPC Toolchain still failing - same problems</li>
489
490   <li>Recreated cvsr.php - file was previously deleted</li>
491 </ul>
492                               
493                               <h3>Bluecat build failing</h3>
494
495 <p>Alexander Sanochkin wrote:</p>
496
497 <blockquote>
498 <p>Andrew,</p>
499
500 <p>It seems you tried to use a build tag which was not ready to rebuild BlueCat at that time. Also please note that the main BC build script has changed due to updating the BC cross compiler to version 3.4.3. The script is called do_it-bc5.0-gcc_3.4.3. You can get it from the BlueCat CVS. (/cm/CVS/BlueCat/eng/int/scripts).</p>
501
502 <p>Regarding the glib build problem we can not provide intelligent comments at this time as it seems that the 20050314 environment
503 is not available for us on the jaguar machine.</p>
504 </blockquote>
505
506 <p>What build tag are we supposed to use? I found R_5_2_1_ppc_20050319 and I assume that is what I should use. However the difference between do_it-bc5.0-gcc_3.4.3 and do_it-bc5.0 is merely:</p>
507
508 <div class="code"><pre>
509 [int@jaguar loc_archive]$ diff do_it-bc5.0.orig do_it-bc5.0-gcc_3.4.3.orig
510 108c108
511 <   export BC_TARGET=$BLUECAT_TARGET_CPU-lynx-linux-bluecat
512 ---
513 >   export BC_TARGET=$BLUECAT_TARGET_CPU-lynx-linux-gnubc
514 </pre></div>
515
516 <p>And, if I might ask, why all the version numbers in the file? Why isn't it just named do_it and depending on which CVS (RCS?) tag you use you get a 5.0 or a 5.0-gcc_3.4.3 version?</p>
517
518 <p>Also, with do_it-bg5.0-gcc_3.4.3 I suspect that the changes I made for the patch-spec (changing do_step to perform patches between steps 1 and 2 when run in stepwise fashion) have not be incorporated. Seems to me that there are two steps being done in automated mode (i.e. doing all steps at one time) that are not performed when doing things in a stepwise fashion. The first is the building of the new GNU Tools which is effectively step 0 and the second is this patching thing which is normally done between steps 1 and 2. Might I suggest that we make these regular steps in their proper order and renumber the rest?</p>
519
520 <p>Build is still failing (on Jaguar). I get to step4 and it fails with:</p>
521
522 <div class="code"><pre>
523 Building glib package step 4.3 at 22:51:22
524 parse_file: build_package failed for glib_trg.spec
525 ---- Step 4 finished successfully at Tue Mar 22 14:51:39 PST 2005 ----
526
527 Looking at step4/build_glib.log I see:
528
529 [int@jaguar step4]$ tail -f build_glib.log
530 + ac_cv_func_getpwuid_r=yes
531 + ac_cv_func_mutex_trylock=yes
532 + ac_cv_func_cond_timedwait=yes
533 + glib_cv_sizeof_gmutex=24
534 + glib_cv_byte_contents_gmutex=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
535 + ./configure --build=i386-linux-gnu --host=ppc-bluecat-linux
536 configure: warning: Could not determine POSIX flag. (-posix didn't work.)
537 configure: error: can not run test program while cross compiling
538 error: Bad exit status from /usr/lynx/loc_archive/build/20050319/var/tmp/rpm-tmp.40621 (%build)
539     Bad exit status from /usr/lynx/loc_archive/build/20050319/var/tmp/rpm-tmp.40621 (%build)
540 </pre></div>
541
542 <p>Attempting to execute rpm-tmp.40621 reveals:</p>
543
544 <div class="code"><pre>
545 gcc -g -O2 -Wall -D_REENTRANT -o testglib testglib.o .libs/libglib.a
546 .libs/libglib.a(gmessages.o): In function `g_logv':
547 /usr/lynx/loc_archive/build/20050319/cdt/src/bluecat/BUILD/glib-1.2.10/gmessages.c:343: undefined reference to `va_copy'
548 .libs/libglib.a(gstrfuncs.o): In function `g_strdup_vprintf':
549     /usr/lynx/loc_archive/build/20050319/cdt/src/bluecat/BUILD/glib-1.2.10/gstrfuncs.c:154: undefined reference to `va_copy'
550 collect2: ld returned 1 exit status
551 make[2]: *** [testglib] Error 1
552 make[2]: Leaving directory `/usr/lynx/loc_archive/build/20050319/cdt/src/bluecat/BUILD/glib-1.2.10'
553 make[1]: *** [all-recursive] Error 1
554 make[1]: Leaving directory `/usr/lynx/loc_archive/build/20050319/cdt/src/bluecat/BUILD/glib-1.2.10'
555 make: *** [all-recursive-am] Error 2
556 + exit 0
557 </pre></div>
558
559 <p>Any ideas?</p>
560
561 <h3>Native PPC Toolchain failure</h3>
562
563 <p>Build of the Native PPC Toolchain keeps failing for me but not for Oleg. Oleg's been suggesting that I lower the ulimits to -s 100000 and -d 200000, which I did but which fails for me and not Oleg. Oleg writes:</p>
564
565 <blockquote>
566 <p>Perhaps the time of the day has some effect on the file system behaviour (or an increased local network activity during the business time has some adverse effect on the system stability). Please try to start the toolchain build at your end-of-business time.</p>
567
568 <p>Meanwhile, we will ponder on what else can be wrong.</p>
569 </blockquote>
570
571                               
572                               <p class="entry-footer">
573                                  <span class="post-footers">Posted by  at  4:44 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000322.html">Permalink</a>
574                                  
575                                  
576                               </p>
577                            </div>
578                         </div>
579                      </div>
580                      
581                      
582
583                      <h2 class="date-header">March 18, 2005</h2>
584                      <a id="a000318"></a>
585                      <div class="entry" id="entry-318">
586                         <h3 class="entry-header">NetSNMP/Native x86 toolchain problems/ECR Linkify</h3>
587                         <div class="entry-content">
588                            <div class="entry-body">
589                               <ul>
590   <li>Change NetSNMP package to automatically handle installation of snmpd-sample.conf file</li>
591
592   <li>Native toolchain build still failing despite assistance from Moscow</li>
593
594   <li>Made ECR Linkify better</li>
595                               
596                               <h3>NetSNMP</h3>
597
598 <p>Changed install_snmp.sh to install snmpd-sample.conf into /usr/local/share/snmp/snmpd.conf for both the bin and src packages. In the bin package we put a copy of snmpd-sample.conf into tmp.</p>
599
600 <p>Still need to figure out how/where to check this into the RCS tree.</p>
601
602 <p>Build scripts do not do this automatically but the build scripts have other issues.</p>
603                               
604                               <p class="entry-footer">
605                                  <span class="post-footers">Posted by  at  5:34 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000318.html">Permalink</a>
606                                  
607                                  
608                               </p>
609                            </div>
610                         </div>
611                      </div>
612                      
613                      
614
615                      <h2 class="date-header">March 17, 2005</h2>
616                      <a id="a000319"></a>
617                      <div class="entry" id="entry-319">
618                         <h3 class="entry-header">Bluecat Build</h3>
619                         <div class="entry-content">
620                            <div class="entry-body">
621                               <ul>
622   <li>Bluecat build's step 4 failed again</li>
623
624   <li>Attempting to reproduce build with old tag</li>
625 </ul>
626                               
627                               <p class="entry-footer">
628                                  <span class="post-footers">Posted by  at  5:35 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000319.html">Permalink</a>
629                                  
630                                  
631                               </p>
632                            </div>
633                         </div>
634                      </div>
635                      
636                      
637
638                      <h2 class="date-header">March 16, 2005</h2>
639                      <a id="a000320"></a>
640                      <div class="entry" id="entry-320">
641                         <h3 class="entry-header">Bluecat Build</h3>
642                         <div class="entry-content">
643                            <div class="entry-body">
644                               <ul>
645   <li>Building Bluecat from start including rebuilding the GNU Tools</li>
646
647   <li>Proceeded through steps 1-4 but I still have problems at step 4</li>
648 </ul>
649                               
650                               <h3>Bluecat GNU Tools</h3>
651
652 <p>There are two "steps" that are not really steps and that are performed only in <i>automated</i> mode. Moscow had reccommended that we insert an exit statement in a certain location and run in automated mode (as root I can only assume) to get the Bluecat GNU Tools. I had originally subsetted this out and built them as int. However since I was having so many odd problems I decided to go strictly as Moscow instructs.</p>
653                               
654                               <p class="entry-footer">
655                                  <span class="post-footers">Posted by  at  5:40 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000320.html">Permalink</a>
656                                  
657                                  
658                               </p>
659                            </div>
660                         </div>
661                      </div>
662                      
663                      
664
665                      <h2 class="date-header">March 15, 2005</h2>
666                      <a id="a000317"></a>
667                      <div class="entry" id="entry-317">
668                         <h3 class="entry-header">BC Step 3 build failure</h3>
669                         <div class="entry-content">
670                            <div class="entry-body">
671                               <ul>
672   <li>Attempting to rebuild BC with new tag.</li>
673
674   <li>Installed RH 8.0 on new machine</li>
675 </ul>
676                               
677                               <h3>Bluecat build failure in step 3</h3>
678
679 <p>I was asked to rebuild Bluecat using R_5_2_1_ppc_20050314 as a tag. I tried doing this but it is failing. I made it to step 4 when it was failing with something about unable to find /arch/ppc/Makefile or something like that. I decided to instead go back to the beginning and make the GNU Tools (step 0) just to be sure. Now I get stuck at step 3. It fails in an odd way too. In /usr/lynx/loc_archive under LOGS I have the following at the tail of the step3 log:</p>
680
681 <div class="code"><pre>
682 Building glibc package step 3.3 at 14:14:07
683 Done
684 Installing glibc package
685 Done
686 Building glib package step 3.4 at 17:06:12
687 parse_file: build_package failed for glib_cdt.spec
688 ---- Step 3 finished successfully at Tue Mar 15 17:07:47 PST 2005 ----
689 +++ Exit .... +++
690 </pre></div>
691
692 <p>However in archive/20050314/ppc/logs/step3 for build_glib.log I have:</p>
693
694 <div class="code"><pre>
695 testglib.c:915: warning: const qualifier ignored on asm
696 .libs/libglib.so: undefined reference to `__ctype_b'
697 .libs/libglib.so: undefined reference to `__ctype_toupper'
698 .libs/libglib.so: undefined reference to `__ctype_tolower'
699 collect2: ld returned 1 exit status
700 make[2]: *** [testglib] Error 1
701 make[1]: *** [all-recursive] Error 1
702 make: *** [all-recursive-am] Error 2
703 error: Bad exit status from /usr/lynx/loc_archive/build/20050314/var/tmp/rpm-tmp.95118 (%build)
704    Bad exit status from /usr/lynx/loc_archive/build/20050314/var/tmp/rpm-tmp.95118 (%build)
705 </pre></div>
706
707 <p>Do you know what is going wrong?</p>
708 <hr>
709 <h4>Later I wrote:</h4>
710
711 <p>Well rebuilding didn't help (I didn't think it would). Things seem to be failing in the rpm-tmp.<num> script. This script seems to be dynamically created. It sets some variables and basically does a configure and make in build/20050314/cdt/src/bluecat/BUILD/glib-1.2.10. This eventually boils down to the following gcc command:</p>
712
713 <div class="code"><pre>
714 [int@jaguar glib-1.2.10]$ gcc -O2 -g -march=i386 -Wall -D_REENTRANT -o .libs/testglib testglib.o .libs/libglib.so -Wl,--rpath -Wl,/cdt/lib
715 .libs/libglib.so: undefined reference to `va_copy'
716 collect2: ld returned 1 exit status
717 </pre></div>
718
719 <p>Oh, BTW, the old machine (penguin) has been renamed to jaguar but is otherwise the same.</p>
720
721 <p>I'm attempting to run step 3 again but I fear it'll just have the same error.</p>
722
723 <p>Note I will be monitoring email and this build from home so if you have any ideas send them right away.</p>
724
725                               
726                               <p class="entry-footer">
727                                  <span class="post-footers">Posted by  at  3:41 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000317.html">Permalink</a>
728                                  
729                                  
730                               </p>
731                            </div>
732                         </div>
733                      </div>
734                      
735                      
736
737                      <h2 class="date-header">March 14, 2005</h2>
738                      <a id="a000316"></a>
739                      <div class="entry" id="entry-316">
740                         <h3 class="entry-header">NetSNMP/Bluecat build/Build Machine</h3>
741                         <div class="entry-content">
742                            <div class="entry-body">
743                               <ul>
744   <li>Rebuilt NetSNMP 5.1.1. Need to package src before build</li>
745
746   <li>Restarted Bluecat build for new label</li>
747
748   <li>Installing RH 8.0 on new machine</li>
749 </ul>
750                               
751                               <h3>NetSNMP 5.1.1</h3>
752
753 <p>We have a chicken and egg situation here. Apparently the src package needs to be built before the build happens, yet it also needs kern_mib.o and kern_mib.d.o, which you get <b>after</b> the build!. The scripts do not handle this gracefully so I rebuilt the packages by hand. Julia says the build is OK now but repeatability will be a problem</p>
754
755 <h3>Installation of RH 8.0</h3>
756
757 <p>After some initial trouble installing from CDROM (Needed to set IDE to Legacy in the BIOS) I managed to install RH 8.0. However I sized the root drive to only 2048 Meg. Turns out the RH 8.0 installation took 2020 Meg so I want to redo this. Upon reinstallation I'm again having problems with the CD. Seems the first CD has etchings on the outermost ring of the CD itself. Not sure if I caused this or if it was there before. May need to get anotehr CD for RH 8.0</p>
758                               
759                               <p class="entry-footer">
760                                  <span class="post-footers">Posted by  at 11:36 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000316.html">Permalink</a>
761                                  
762                                  
763                               </p>
764                            </div>
765                         </div>
766                      </div>
767                      
768                      
769
770                      <h2 class="date-header">March 11, 2005</h2>
771                      <a id="a000315"></a>
772                      <div class="entry" id="entry-315">
773                         <h3 class="entry-header">NetSNMP</h3>
774                         <div class="entry-content">
775                            <div class="entry-body">
776                               <ul>
777   <li>Rebuilt NetSNMP.</li>
778 </ul>
779                               
780                               <p class="entry-footer">
781                                  <span class="post-footers">Posted by  at  4:59 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000315.html">Permalink</a>
782                                  
783                                  
784                               </p>
785                            </div>
786                         </div>
787                      </div>
788                      
789                      
790
791                      <h2 class="date-header">March 10, 2005</h2>
792                      <a id="a000314"></a>
793                      <div class="entry" id="entry-314">
794                         <h3 class="entry-header">Native x86 toolchain/Bluecat Installation/ECRD timeouts</h3>
795                         <div class="entry-content">
796                            <div class="entry-body">
797                               <ul>
798   <li>Discovered problems with Native x86 toolchain.</li>
799
800   <li>Attempting Bluecat installation from CDs - taking a long time!</li>
801
802   <li>Implemented timeouts correctly in ecrd, which was having client deadlocks every night</li>
803 </ul>
804                               
805                               <h3>Native X86 Toolchain Build Problem</h3>
806
807 <p>When building the toolchain on x86 natively I kept getting different failures. It was frustrating to say the least. Conferred with Adam and he remembered a problem with the kernel changing timestamps on files due to an nmap problem (ECR <a href="http://saturn/ecr/ecr.php?ecr=22905">22905</a>) which still seems to be a problem. Emailed Vlad about this...</p>
808
809 <h3>Bluecat installation from CDs</h3>
810
811 <p>Bluecat installation was takng a long time due to timeouts in NFS for a stale file handle for some remotely mounted CDROM. Still have an issue with one RPM that has int:staff ownership:</p>
812
813 <div class="code><pre>
814     Installing BlueCat_ppc/cdt/RPMS/libgcc_cdt-3.2.2-1.i386.rpm
815     error: failed to stat /mnt/qa: Stale NFS file handle
816     warning: user int does not exist - using root
817     warning: group staff does not exist - using root                              
818     warning: user int does not exist - using root
819     warning: group staff does not exist - using root
820     warning: user int does not exist - using root
821     warning: group staff does not exist - using root
822     warning: user int does not exist - using root
823     warning: group staff does not exist - using root
824     warning: user int does not exist - using root
825     warning: group staff does not exist - using root
826     warning: user int does not exist - using root
827     warning: group staff does not exist - using root
828 </pre></div>
829
830 <p>Note that this is the only package that this occurred on. It seems to verify Sasha's assertion that if the rpm's are not packaged by a native user then such warnings will occur. Still it seems odd that these "ppc" rpm's have "i386" in them...</p>
831                               
832                               <p class="entry-footer">
833                                  <span class="post-footers">Posted by  at  4:21 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000314.html">Permalink</a>
834                                  
835                                  
836                               </p>
837                            </div>
838                         </div>
839                      </div>
840                      
841                      
842
843                      <h2 class="date-header">March  9, 2005</h2>
844                      <a id="a000313"></a>
845                      <div class="entry" id="entry-313">
846                         <h3 class="entry-header">PPC Toolchain finally builds!</h3>
847                         <div class="entry-content">
848                            <div class="entry-body">
849                               <ul>
850   <li>Finally managed to build Native PPC Toolchain after many interruptions due to machine moves</li>
851
852   <li>Revisiting build of Native X86 Toolchain which is failing.</li>
853 </ul>
854                               
855                               <p class="entry-footer">
856                                  <span class="post-footers">Posted by  at  4:11 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000313.html">Permalink</a>
857                                  
858                                  
859                               </p>
860                            </div>
861                         </div>
862                      </div>
863                      
864                      
865
866                      <h2 class="date-header">March  8, 2005</h2>
867                      <a id="a000311"></a>
868                      <div class="entry" id="entry-311">
869                         <h3 class="entry-header">netsnmp/t3 mount/PPC Toolchain</h3>
870                         <div class="entry-content">
871                            <div class="entry-body">
872                               <ul>
873   <li>Build NetSNMP 5.1.1</li>
874
875   <li>Had Jeff export t3:/usr/lynx/archive/ecr so that ECRDig can gain access to auxilary files that people store there. Currently cat.php doesn't work with all the file types it can hit there.</li>
876
877   <li>Figured out that the tar images were not being recreated due to int_tools getting stuck in the demos. Recreated PPC tar image and unpacked it onto target machine.</li>
878 </ul>
879                               
880                               <p class="entry-footer">
881                                  <span class="post-footers">Posted by  at  5:03 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000311.html">Permalink</a>
882                                  
883                                  
884                               </p>
885                            </div>
886                         </div>
887                      </div>
888                      
889                      
890
891                      <h2 class="date-header">March  7, 2005</h2>
892                      <a id="a000310"></a>
893                      <div class="entry" id="entry-310">
894                         <h3 class="entry-header">Bluecat CDs</h3>
895                         <div class="entry-content">
896                            <div class="entry-body">
897                               <ul>
898   <li>Burned 4 CDs for Bluecat</ii>
899
900   <li>Continuing to hunt down problem with PPC Toolchain build. Missed a file. After including it libc.a does have atoi but it doesn't seem to be being packaged into the product image!</li>
901 </ul>
902                               
903                               <p class="entry-footer">
904                                  <span class="post-footers">Posted by  at 10:12 AM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000310.html">Permalink</a>
905                                  
906                                  
907                               </p>
908                            </div>
909                         </div>
910                      </div>
911                      
912                      
913
914                      <h2 class="date-header">March  4, 2005</h2>
915                      <a id="a000309"></a>
916                      <div class="entry" id="entry-309">
917                         <h3 class="entry-header">Yet more building</h3>
918                         <div class="entry-content">
919                            <div class="entry-body">
920                               <ul>
921   <li>Finished Bluecat build process!</li>
922
923   <li>Pulling ECR 23001 into the LynxOS build for eventual Native PPC Toolchain build</li>
924 </ul>
925                               
926                               <p>Regarding ECR 23001:</p>
927
928 <p>I was trying to build the native PPC toolchain again. The ECR I submitted was 23184 but that was dupped to 22979 which is in Pending Review state. So I pulled 22979 and attempted to build. But 22979 depends on 23001 (also in Pending Review). Is it OK to pull 23001 and continue onward? How are dependencies between ECRs normally handled?</p>
929
930 <blockquote>
931   <p><b>Note:</b> I say "depends on" but I'm not sure that that is the right terminology. I don't really believe that 23001 depends on 22979 nor vica versa rather there doesn't seem to be a clear process here. 22979 pulls in src/lib/libc/Makefile revision 10.25 but 10.24 has a change for 23001 in it. So 23001 is needed because of this. The other files involved in 23001 do not have the Lion_lynxos_012405 tag so they are not picked up in the normal build process - hence the build fails. This is not a classic dependency rather it's an overrunning of checkins. Of course, maybe my analysis of this situation is flawed.</p>
932 </blockquote>
933                               
934                               <p class="entry-footer">
935                                  <span class="post-footers">Posted by  at  1:53 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000309.html">Permalink</a>
936                                  
937                                  
938                               </p>
939                            </div>
940                         </div>
941                      </div>
942                      
943                      
944
945                      <h2 class="date-header">March  2, 2005</h2>
946                      <a id="a000308"></a>
947                      <div class="entry" id="entry-308">
948                         <h3 class="entry-header">Building, building...</h3>
949                         <div class="entry-content">
950                            <div class="entry-body">
951                               <ul>
952   <li>Pulled ECR 22979 and rebuilt lynxos</li>
953
954   <li>Placed new LynxOS on T3 and updated t-mcpn765-1</li>
955
956   <li>Rebuild of Toolchain still fails with unresolved references to atoi!</li>
957
958   <li>Found some corruption in LynxOS CVS. Working to get a list of files and to have this fixed</li>
959
960   <li>Bluecat build, step 4 still failing...</li>
961
962   <li>Attempting to build TOB Toolchain on x86 still failing</li>
963 </ul>
964                               
965                               <p class="entry-footer">
966                                  <span class="post-footers">Posted by  at  6:26 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000308.html">Permalink</a>
967                                  
968                                  
969                               </p>
970                            </div>
971                         </div>
972                      </div>
973                      
974                      
975
976                      <h2 class="date-header">March  1, 2005</h2>
977                      <a id="a000307"></a>
978                      <div class="entry" id="entry-307">
979                         <h3 class="entry-header">More BC building</h3>
980                         <div class="entry-content">
981                            <div class="entry-body">
982                               <ul>
983   <li>Moscow's BC build failed in step 4 due to lack of disk space. Moved guest account over to new partition</li>
984
985   <li>Step 4 is not building for me either. Emailed Moscow</li>
986
987   <li>Returning to other builds. PPC Native Toolchain build was failing - attempting to resolve that. Also X86 Native Toolchain build also failing.</li>
988
989   <li>Several enhancements to ECR Dig:</li>
990
991     <ul>
992       <li>ECRs now include status line with Status, State and Severity</li>
993
994       <li>Http and ftp references are made into links</li>
995
996       <li>ECR numbers are now made into links</li>
997
998       <li>Added footer with "Back to ECR Dig" link</li>
999     </ul>
1000 </ul>
1001                               
1002                               <p class="entry-footer">
1003                                  <span class="post-footers">Posted by  at  4:41 PM</span> <span class="separator">|</span> <a class="permalink" href="http://defaria.com/blogs/Status/archives/000307.html">Permalink</a>
1004                                  
1005                                  
1006                               </p>
1007                            </div>
1008                         </div>
1009                      </div>
1010                      
1011                   </div>
1012                </div>
1013             </div>
1014          </div>
1015       </div>
1016    </div>
1017 </body>
1018 </html>