Opened 9 years ago
Closed 9 years ago
#6763 closed enhancement (fixed)
gnumeric-1.12.23
| Reported by: | Fernando de Oliveira | Owned by: | Douglas R. Reno |
|---|---|---|---|
| Priority: | normal | Milestone: | 7.8 |
| Component: | BOOK | Version: | SVN |
| Severity: | normal | Keywords: | |
| Cc: |
Description (last modified by )
http://ftp.gnome.org/pub/gnome/sources/gnumeric/1.12/gnumeric-1.12.23.tar.xz
MD5: 4e1844f3fd6162a7b0a96d3325dcac5d
http://ftp.gnome.org/pub/gnome/sources/gnumeric/1.12/gnumeric-1.12.23.sha256sum
00474cbf1e70f62062974d0beb5f65ebc45d1047bc8fd0a1133e3d9725e19551
http://ftp.gnome.org/pub/gnome/sources/gnumeric/1.12/gnumeric-1.12.23.news
Gnumeric 1.12.23 Andreas: * Fix memory handling error on fuzzed sxc file. [#748535] * Improve import/export of page layout from/to ODF. * Improve function import to ODF. [#750627] * Add CONFIDENCE.T, MODE.MULT, PERCENTILE.EXC, PERCENTRANK.EXC and QUARTILE.EXC. * Fuzzed file fixes. [#751060] [#751501] [#751851] [#751922] * Fix ODF import of named expressions. [#750627] * Import reference unions from ODF. [#750627] * Improve CONVERT function. Jean: * Fix xlsx import of plot area manual layout. [#748016] * Fix out of bounds read. [#749121] * Fuzzed file fixes. [#750042] [#751217] [#751270] [#751271] [#751383] [#751384] [#751758] [#751744] [#751908] [#751920] [#751249] [#751945] [#751909] [#751946] [#751968] [#751969] [#751971] [#751989] Morten: * Fix import/export of graph backplane. * Fix export of unlabelled axes. * Fix export of rotated axis labels. * Fix xlsx save crash related to shared strings. [#748477] * Solver code refactoring. * Plug leaks. * Fuzzed file fixes. [#748595] [#748597] [#749031] [#749030] [#749069] [#748533] [#749118] [#749166] [#749181] [#749184] [#749236] [#749240] [#749234] [#749235] [#749271] [#749270] [#749424] [#749917] [#749919] [#750043] [#750044] [#750046] [#750811] [#750810] [#750857] [#750864] [#750862] [#750858] [#751126] [#751254] [#751253] [#750851] [#751258] [#751259] [#751502] [#751390] [#751579] [#751659] [#751660] [#751662] [#751970] [#752022] [#751988] [#752021] [#752080] [#752081] [#752124] * Make solver check linearity of model. * Fix xls saving of marker style. [#749185] * Make compilation with clang work again. [#749138] * Fix xlsx problem with import of header/footer. [#750853] * Fix xls save problem with characters outside codepage. [#750856] * Fix named expression problem. [#751056] * Fix xlsx namespace problem. [#751120] * Protect database functions against malformed database. [#751392] * Fix csv problem with invalid UTF-8 data. [#751577] * Fix GROWTH. [#751658] * Restrict TABLE to array context. [#751871] --------------------------------------------------------------------------
Change History (23)
comment:1 by , 9 years ago
| Owner: | changed from to |
|---|---|
| Status: | new → assigned |
comment:2 by , 9 years ago
comment:3 by , 9 years ago
Re: requested by Doug: (probably also this one - libgsf-1.14.34) goffice-0.10.23 and gnumeric-1.12.23
Yes, please.
If you wish, take also libgsf, required by goffice (and wv, but I verified it is OK, so, no need to check wv).
comment:4 by , 9 years ago
| Owner: | changed from to |
|---|---|
| Status: | assigned → new |
comment:5 by , 9 years ago
| Status: | new → assigned |
|---|
comment:6 by , 9 years ago
Doug, since last version, tests are OK, with me.
Perhaps it is time to modify the tests text, if you have similar result.
Probably Pierre might have left, because I did have problems in the past.
Pierre, please, would confirm that?
There is a large amount of SKIP, but they seem related to missing dependencies.
comment:7 by , 9 years ago
IIRC, quite a few testfiles (spreadsheests in various formats) are missing in the tarball, and can be obtained from upstream repo. I asked on their list why it was so, and was answered that tests were for developpers, not for users...
My last test log is for version 1.12.21 and shows:
TOTAL: 118 PASS: 72 SKIP: 45 XFAIL: 0 FAIL: 1 XPASS: 0 ERROR: 0
Some tests use valgrind, so I guess they are skipped if valgrind is not found.
comment:8 by , 9 years ago
Forgot to answer the original question, sorry. I do not remember whether I changed the text for tests. I think I estimated that the one failure is compatible with "some tests may fail". Also, ISTR that I added the dconf dependency. Without it, there are a lot of fake errors, due to the fact that gnumeric sends an alert message when it does not find the dconf database, which is interpreted by the ttest harness as a wrong output.
comment:9 by , 9 years ago
Thanks for the explanations and info, Pierre!
I don't have the results for version 1.12.21.
For versions 1.12.22 and 1.12.23, I have:
# TOTAL: 124 # PASS: 80 # SKIP: 44 # XFAIL: 0 # FAIL: 0 # XPASS: 0 # ERROR: 0
Therefore, either I have some dependency you didn't, or the tests have improved.
Anyway, when I get test results different from what is in the book, sometimes I leave what is in the book, other times, I add a sentence.
If this time Doug has no failure, perhaps we could remove some parts, and the -k?
follow-up: 12 comment:10 by , 9 years ago
Does anyone build Yelp for viewing the help? It looks like as though I would have to build WebKitGTK+-2.4.9, so I was wondering whether or not you guys used it when compiling Gnumeric. Fernando, do you already have that installed?
follow-up: 13 comment:11 by , 9 years ago
Fernando, do you have Valgrind installed on your box? That is the only difference that I can think of, besides the fact that I am running an i686 PC and that I do not have Yelp installed at this time.
# TOTAL: 124 # PASS: 5 # SKIP: 48 # XFAIL: 0 # FAIL: 71 # XPASS: 0 # ERROR: 0
Here are the individual tests:
FAIL: t1000-statfuns.pl FAIL: t1001-mathfuns.pl FAIL: t1002-finfuns.pl FAIL: t1003-engfuns.pl FAIL: t1004-textfuns.pl FAIL: t1005-logfuns.pl FAIL: t1006-datefuns.pl FAIL: t1007-dbfuns.pl FAIL: t1008-infofuns.pl FAIL: t1009-lookfuns.pl FAIL: t1010-bitwise.pl FAIL: t1011-yalta2008.pl FAIL: t1012-burkhardt.pl FAIL: t1013-crlibm.pl SKIP: t1100-chitest.pl SKIP: t1101-ftest.pl SKIP: t1102-ttest.pl SKIP: t1103-sumif.pl SKIP: t1104-linest.pl FAIL: t1105-gamma.pl FAIL: t1800-operators.pl SKIP: t1900-array.pl FAIL: t1901-regress.pl SKIP: t1902-array-intersection.pl FAIL: t2000-names-insdel.pl FAIL: t2001-func-help.pl FAIL: t2002-nonascii-numbers.pl FAIL: t2003-random-generators.pl FAIL: t2004-insdel-colrow.pl FAIL: t2800-style-optimizer.pl SKIP: t5900-sc.pl SKIP: t5901-qpro.pl SKIP: t5902-applix.pl SKIP: t5903-sylk.pl SKIP: t5904-mps.pl FAIL: t5905-guppi.pl FAIL: t6000-statfuns.pl FAIL: t6001-mathfuns.pl FAIL: t6002-objs.pl FAIL: t6003-lookfuns.pl FAIL: t6004-finfuns.pl SKIP: t6050-vba-roundtrip.pl SKIP: t6100-statfuns-ods.pl SKIP: t6101-mathfuns-ods.pl SKIP: t6102-objs-ods.pl SKIP: t6103-lookfuns-ods.pl SKIP: t6104-finfuns-ods.pl FAIL: t6150-ods-syntax.pl FAIL: t6151-gnumeric-syntax.pl FAIL: t6152-xlsx-syntax.pl FAIL: t6160-ods-deterministic.pl FAIL: t6161-xlsx-deterministic.pl FAIL: t6500-strings.pl FAIL: t6501-numbers.pl FAIL: t6502-styles.pl FAIL: t6503-page-setup.pl FAIL: t6504-formula.pl FAIL: t6505-colrow.pl FAIL: t6506-cell-comments.pl FAIL: t6507-split-panes.pl FAIL: t6508-sheet-formatting.pl FAIL: t6509-validation.pl FAIL: t6510-auto-filter.pl FAIL: t6511-solver.pl FAIL: t6512-format.pl FAIL: t6513-merges.pl FAIL: t6514-rich-text.pl FAIL: t6515-cond-format.pl FAIL: t6516-graph.pl FAIL: t6517-names.pl FAIL: t6518-objects.pl FAIL: t6519-selection.pl FAIL: t6520-sheet-tab.pl FAIL: t7000-goal-seek.pl SKIP: t7100-solver-blend.pl SKIP: t7101-solver-afiro.pl SKIP: t7110-solver-Misra1a.pl SKIP: t7111-solver-Chwirut2.pl SKIP: t7112-solver-Chwirut1.pl SKIP: t7113-solver-Lanczos3.pl SKIP: t7114-solver-Gauss1.pl SKIP: t7115-solver-Gauss2.pl SKIP: t7116-solver-DanWood.pl SKIP: t7117-solver-Misra1b.pl SKIP: t7120-solver-Kirby2.pl SKIP: t7121-solver-Hahn1.pl SKIP: t7122-solver-Nelson.pl SKIP: t7123-solver-MGH17.pl SKIP: t7124-solver-Lanczos1.pl SKIP: t7125-solver-Lanczos2.pl SKIP: t7126-solver-Gauss3.pl SKIP: t7127-solver-Misra1c.pl SKIP: t7128-solver-Misra1d.pl SKIP: t7129-solver-Roszman1.pl SKIP: t7130-solver-ENSO.pl SKIP: t7140-solver-MGH09.pl SKIP: t7141-solver-Thurber.pl SKIP: t7142-solver-BoxBOD.pl SKIP: t7143-solver-Rat42.pl SKIP: t7144-solver-MGH10.pl SKIP: t7145-solver-Eckerle4.pl SKIP: t7146-solver-Rat43.pl SKIP: t7147-solver-Bennett5.pl FAIL: t8000-valgrind.pl FAIL: t8001-valgrind-xls.pl FAIL: t8002-valgrind-gnumeric.pl FAIL: t8003-valgrind-pdf.pl FAIL: t8004-valgrind-xls-macro.pl FAIL: t8005-valgrind-chart-fit.pl FAIL: t8010-valgrind-ods.pl FAIL: t8011-valgrind-xlsx.pl FAIL: t8050-valgrind-sstest.pl PASS: t8100-null-false.pl PASS: t8101-gfree.pl SKIP: t8102-glade-visible.pl PASS: t8103-gtk.pl PASS: t8104-glade-parsing.pl FAIL: t8200-gslice1.pl FAIL: t8201-gslice2.pl FAIL: t8202-gslice3.pl FAIL: t9000-ssindex.pl FAIL: t9001-ssconvert-resize.pl FAIL: t9100-number-match.pl PASS: t9999-epilogue.pl
I will wait for feedback before I continue. Those results are awful!!!!!
comment:12 by , 9 years ago
Replying to renodr:
Does anyone build Yelp for viewing the help?
Yes, but not for gnumeric (or Abiword, IIRC).
It looks like as though I would have to build WebKitGTK+-2.4.9, so I was wondering whether or not you guys used it when compiling Gnumeric.
Not for Gnumeric.
Fernando, do you already have that installed?
Yes, but for the Gnomes.
comment:13 by , 9 years ago
Replying to renodr:
Fernando, do you have Valgrind installed on your box? That is the only difference that I can think of, besides the fact that I am running an i686 PC and that I do not have Yelp installed at this time.
...
I will wait for feedback before I continue. Those results are awful!!!!!
Yes. I do. Please, build it also for libgsf.
I am not reading the book now, but if valgrind is not dependency listed, I think it should, in both packages, for the tests, without it, many failures .
comment:14 by , 9 years ago
OK, I'm not the on to be followed, regarding dependencies.
But probably in both, libgsf and Gnumeric, I have the impression that valgrind should be recommended (if running the tests), even required, or something like that.
follow-up: 17 comment:15 by , 9 years ago
I had valgrind installed before I did Gnumeric, thats why I was asking if you had it or not. I wonder what is causing the issue with the tests suites. libgsf and goffice were both fine, its just Gnumeric that isn't testing properly. I did notice a mention of an external set of files for test suite, correct? Where can I find those?
follow-up: 18 comment:16 by , 9 years ago
I just thought of something... I am doing a DESTDIR install for the book update, and then I normally install it into /usr. When I get home, I will install Gnumeric into /usr and see if that fixes it or not.
comment:17 by , 9 years ago
Replying to renodr:
I did notice a mention of an external set of files for test suite, correct? Where can I find those?
I never used them.
comment:7 Changed 25 hours ago by pierre.labastie
External files. I remember him mentioning it before. Either in some ticket or in -dev list.
comment:8 Changed 25 hours ago by pierre.labastie
Pierre mentions dconf. Don't know if it is in the page. Might be necessary to add for tests, at lest, if it is not.
follow-up: 19 comment:18 by , 9 years ago
Replying to renodr:
I just thought of something... I am doing a DESTDIR install for the book update, and then I normally install it into /usr. When I get home, I will install Gnumeric into /usr and see if that fixes it or not.
I didn't understand completely. If I do a make DESTDIR... install, I may use the same source and do a make install. Usually, if it a short SBU package, instead, I do a second build from scratch for definitive install. Second is also used to check that it supports parallel build.
Never copy or install the DESTDIR files. Had many problems in the past and in the book you see some notes about fixes, if you do that.
comment:19 by , 9 years ago
Replying to fo:
Replying to renodr:
I just thought of something... I am doing a DESTDIR install for the book update, and then I normally install it into /usr. When I get home, I will install Gnumeric into /usr and see if that fixes it or not.
I didn't understand completely. If I do a make DESTDIR... install, I may use the same source and do a make install. Usually, if it a short SBU package, instead, I do a second build from scratch for definitive install. Second is also used to check that it supports parallel build.
Never copy or install the DESTDIR files. Had many problems in the past and in the book you see some notes about fixes, if you do that.
I had not planned on copying or installing the DESTDIR files. I will do a make install instead and then try running the tests. Hopefully that can give me a definitive answer on whether or not the tests work. I also am going to rebuild afterwards, since my scripts will only output the size of the package and such if the script completed without error.
I also already installed DCONF, since that and valgrind were listed in the book as dependencies for the tests.
comment:20 by , 9 years ago
These are my new test results after running "make install":
FAIL: t6500-strings.pl FAIL: t6506-cell-comments.pl FAIL: t6509-validation.pl FAIL: t8000-valgrind.pl FAIL: t8001-valgrind-xls.pl FAIL: t8002-valgrind-gnumeric.pl FAIL: t8003-valgrind-pdf.pl FAIL: t8004-valgrind-xls-macro.pl FAIL: t8005-valgrind-chart-fit.pl FAIL: t8010-valgrind-ods.pl FAIL: t8011-valgrind-xlsx.pl FAIL: t8050-valgrind-sstest.pl
# TOTAL: 124 # PASS: 69 # SKIP: 43 # XFAIL: 0 # FAIL: 12 # XPASS: 0 # ERROR: 0
Note that the valgrind tests still fail even though I know that I have valgrind installed. I have ran Gnumeric and it seems to run fine. Should I proceed with the update now?
comment:21 by , 9 years ago
Well, I think you can do the update.
Independently from that, it would be good to understand the reason for those problems, exactly with valgrind. Can you find in the web references to these kind of problems?
I have:
$ valgrind --version valgrind-3.10.1
I'm sorry, but I don't know how to help with this problem.
Eyes too tired at this hour.
comment:22 by , 9 years ago
| Description: | modified (diff) |
|---|
Can I take this?