From: Martin Q. <mar...@lo...> - 2005-02-23 23:18:04
|
Hello, make test fails on my box. Is it just me or the lastest changes to allow several versions to coexist broke it? As usual, I barely understand what I see, and I may well be wrong. :) Thanks, Mt. |
From: Dowling, W. (TS USA) <wil...@th...> - 2005-02-24 20:44:21
|
On Wed, 2005-02-23 at 18:17, Martin Quinson wrote: > Hello, > > make test fails on my box. Is it just me or the lastest changes to allow > several versions to coexist broke it? [...] I am seeing this: $ make BASE_DIR=/tmp/foo test Testing... make -C examples test make[1]: Entering directory `/space1/proj-temp/flexml/flexml/examples' Testing... ../flexml.pl -vD -a my-show.act my.dtd FleXML version Id: flexml.pl,v 1.44 2005/02/23 23:08:16 mquinson Exp. Fetching URL <file:my.dtd>. Processing DTD in my.dtd. Post-processing DTD. Writing XML application onto `my-show.c'. Can't exec "./flexml-act": No such file or directory at ../flexml.pl line 1527. readline() on closed filehandle ACTIONS at ../flexml.pl line 1534. ../flexml.pl: Cannot close pipe to flexml-act: Bad file descriptor make[1]: *** [my-show.c] Error 9 make[1]: Leaving directory `/space1/proj-temp/flexml/flexml/examples' make: *** [test] Error 2 I don't understand why this Makefile runs ../flexml.pl. We should be testing what we just built, right? If that is so, replace all "../flexml.pl" by "../$(FLEXML)" in examples/Makefile, and you get a lot farther. The next place it fails is at Fetching URL <file:xhtml-lat1.ent>. "xhtml1-transitional.dtd", line 29: URL <file:xhtml-lat1.ent> could not be fetched. make[1]: *** [xhtml-href.c] Error 2 make[1]: Leaving directory `/space1/proj-temp/flexml/flexml/examples' make: *** [test] Error 2 Are we missing a file xhtml-lat1.ent? I don't know if it was the latest changes -- I don't normally run 'make test' -- so I don't know when it started failing. The good news is this: When I run my own tests (not the examples/ directory), everything seems to be working. -- William F. Dowling Thomson/ISI (www.thomsonisi.com) 215-386-0100 x-1156 ______________________________________________________________________ This email has been scanned by the MessageLabs Email Security System. For more information please visit http://www.messagelabs.com/email ______________________________________________________________________ |
From: Martin Q. <mar...@lo...> - 2005-03-05 11:36:07
|
On Thu, Feb 24, 2005 at 03:42:20PM -0500, Dowling, Will (TS USA) wrote: > On Wed, 2005-02-23 at 18:17, Martin Quinson wrote: > > Hello, > >=20 > > make test fails on my box. Is it just me or the lastest changes to allow > > several versions to coexist broke it? [...] >=20 > I am seeing this: >=20 > $ make BASE_DIR=3D/tmp/foo test =20 > Testing... > make -C examples test > make[1]: Entering directory `/space1/proj-temp/flexml/flexml/examples' > Testing... > ../flexml.pl -vD -a my-show.act my.dtd > FleXML version Id: flexml.pl,v 1.44 2005/02/23 23:08:16 mquinson Exp. > Fetching URL <file:my.dtd>. > Processing DTD in my.dtd. > Post-processing DTD. > Writing XML application onto `my-show.c'. > Can't exec "./flexml-act": No such file or directory at ../flexml.pl > line 1527. > readline() on closed filehandle ACTIONS at ../flexml.pl line 1534. > ../flexml.pl: Cannot close pipe to flexml-act: Bad file descriptor > make[1]: *** [my-show.c] Error 9 > make[1]: Leaving directory `/space1/proj-temp/flexml/flexml/examples' > make: *** [test] Error 2 >=20 > I don't understand why this Makefile runs ../flexml.pl. We should be > testing what we just built, right? If that is so, replace all > "../flexml.pl" by "../$(FLEXML)" in examples/Makefile, and you get a lot > farther. =20 >=20 > The next place it fails is at >=20 > Fetching URL <file:xhtml-lat1.ent>. > "xhtml1-transitional.dtd", line 29: URL <file:xhtml-lat1.ent> could not > be fetched. > make[1]: *** [xhtml-href.c] Error 2 > make[1]: Leaving directory `/space1/proj-temp/flexml/flexml/examples' > make: *** [test] Error 2 >=20 > Are we missing a file xhtml-lat1.ent? >=20 > I don't know if it was the latest changes -- I don't normally run 'make > test' -- so I don't know when it started failing. >=20 > The good news is this: When I run my own tests (not the examples/ > directory), everything seems to be working. Hello, I'm completly lost here. I'm tweaking the makefile back and forth to try to get it testing something, in vain. If you have any solution leading to even a minimalist step forward, please commit it. I was even unable to do the basic step you describes. Ie, when changing all =2E./flexml-ml occurence, the message I get is not at all the one you give. I'm also puzzeled by the "MAIN PROGRAM" rules in example/Makefile. Do we build a specific version of flexml for testing (and if so, who calls it) or is this a leftover? I consider the lack of testing as a showstopper for the next release (I'd like to push out of the door since in brings some neat stuff such as stack consumtion reduction). Thanks for any help, Mt. |
From: William F. D. <wil...@th...> - 2005-03-18 19:44:03
|
On Sat, 2005-03-05 at 12:35 +0100, Martin Quinson wrote: > Hello, > > I'm completly lost here. I'm tweaking the makefile back and forth to try to > get it testing something, in vain. If you have any solution leading to even > a minimalist step forward, please commit it. I have the "joke" tests working from the Makefile (though not with debugging turned on, which for me causes a core dump in the print_yy_stack() function). But the larger test, which tries to parse FleXML.html using http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd is quite a failure. The problem is that when shack checked in changes to content of the the documentation in FleXML.html, the file became no longer XML. I will separate out the last revision of FleXML.html that was also XML, into a different file, and make sure the test uses that for input. I'm not quite ready to throw out examples/* and replace it with some real tests, but that is what needs to be done. Will -- William F Dowling wil...@th... www.isinet.com |
From: William F. D. <wil...@th...> - 2005-03-18 22:15:29
|
I have checked in changes so that make SUFF= ; cd examples; make should work (i.e. first three tests print "OK" and last two print out sensible results.) I'm still getting a ton of warnings, the whole test Makefile needs to be rethought, but at least now it is minimally working. I'm not happy with the SUFF= bit being necessary for the tests to run. Let me know if you have problems with this. Will -- William F Dowling wil...@th... www.isinet.com |
From: William F. D. <wil...@th...> - 2005-03-21 16:48:56
|
It *should* have worked. The correct instructions are make SUFF= ; cd examples; make SUFF= I have checked in the files xhtml-{special,symbol,lat1}.ent so now the test actually does work. Thanks for your patience. Will On Fri, 2005-03-18 at 17:15 -0500, William F. Dowling wrote: > I have checked in changes so that > make SUFF= ; cd examples; make > should work (i.e. first three tests print "OK" and last two print out > sensible results.) I'm still getting a ton of warnings, the whole test > Makefile needs to be rethought, but at least now it is minimally > working. > > I'm not happy with the SUFF= bit being necessary for the tests to run. > Let me know if you have problems with this. > > Will > > -- William F Dowling wil...@th... www.isinet.com |
From: Martin Q. <mar...@lo...> - 2005-04-06 12:07:13
|
I finally found some time to devote to the project. Your changes wrt to testing infrastructure work for me. And since I had a minor bug against the debian package (packaging issue), I decided to release 1.3 and move the development forward to 1.4 (or 1.3.1 if I did something wrong). I'll announce 1.3 right after this mail. Concerning the testing infrastructure, I think that the easiest could be to make SUFF=3D the default, and let people wanting to use SUFF=3D-$(ver) specifying it from the command line. Want do you think about it? Thanks again for your work,=20 Mt. On Mon, Mar 21, 2005 at 11:46:54AM -0500, William F. Dowling wrote: > It *should* have worked. The correct instructions are > make SUFF=3D ; cd examples; make SUFF=3D >=20 > I have checked in the files xhtml-{special,symbol,lat1}.ent so now the > test actually does work. Thanks for your patience. >=20 > Will >=20 > On Fri, 2005-03-18 at 17:15 -0500, William F. Dowling wrote: > > I have checked in changes so that > > make SUFF=3D ; cd examples; make > > should work (i.e. first three tests print "OK" and last two print out > > sensible results.) I'm still getting a ton of warnings, the whole test > > Makefile needs to be rethought, but at least now it is minimally > > working. > >=20 > > I'm not happy with the SUFF=3D bit being necessary for the tests to run. > > Let me know if you have problems with this. |