Hello,
I'm having difficulty in getting this aspect of PCT working as I would expect. I have a PCT ANT build script which generates a library from an RCODE folder;
build/rcode
build/library
when the script runs the rcode in the RCODE folder is copied and flattened (absolute paths removed) into a temp folder.
build/rcode/temp
It's from this temp folder that the library is built.
My problem is that when the build script compares the new RCODE with the previous library, every file is reported as being a different MD5 and copied to my compare folder (see below).
When I manually checked the md5 (RCODE-INFO:MD5-VALUE) of the library and rcode .r files they are the same!
<!-- Compare RCODE for patch build -->
<copy todir="${prop.BUILD_RCODE}\compare" includeEmptyDirs="false">
<fileset dir="${prop.BUILD_RCODE}\temp" includes="**/*.r">
<RCodeSelector lib="${prop.BUILD_LIBRARY}\${prop.build-name}.pl" mode="md5" />
</fileset>
</copy>
Thanks for any help, it's always much appreciated! :-)
Mark.
Hi,
Please run PCT with -v, that will display the reason why files are selected or not in your RCodeSelector. Otherwise,
PCT Release 202 ( github.com/.../v202 ) includes prolib.jar, a command line utility to display the content of PL files (including MD5). Command line is 'java -jar prolib.jar list -lib /path/to/lib.pl'. That may help you.
Hi,
I ran my build with -v as suggested and piped the output to a txt file;
ant -v build > log.txt
It looks as though some files are being correctly identified whilst others are not!? So, ColumnTypeDelete.r has been identified as having no MD5 changes and has not been copied from temp to compare which is correct. However, ColumnTypeCollection.r has been identified as having a MD5 mis-match and been copied to compare which is incorrect as the MD5is the same.
PLTarget ColumnTypeBrowse.r is an invalid rcode -- Resetting to invalid mark
PLTarget ColumnTypeCollection.r is an invalid rcode -- Resetting to invalid mark
MD5 ColumnTypeCreate.r File1 B8E7CF4CFA6BC5413DB6289E91A4C10A File2 B8E7CF4CFA6BC5413DB6289E91A4C10A
MD5 ColumnTypeDelete.r File1 38F258FA8A8F47B128AE125F8FB014C5 File2 38F258FA8A8F47B128AE125F8FB014C5
---------------------------
Information
---------------------------
Attribute: 1st 2nd RCODE-INFO
CODEPAGE: ISO8859-1 ISO8859-1
CRC-VALUE: 29519 29519
DB-REFS:
FILE-NAME: \rcode\compare\ColumnTypeCollection.r \build\library\ColumnTypeCollection.r
HANDLE: 25 25
INST-PROC: ? ?
IS-CLASS: no no
LANGUAGES: <unnamed> <unnamed>
MD5: B8B31BC73943FACA027846A495F78834 B8B31BC73943FACA027846A495F78834
TBL-CRCS:
TABLES:
TYPE: PSEUDO-WIDGET PSEUDO-WIDGET
MATCH: yes
---------------------------
OK
---------------------------
Thanks.
I've re-edited this response as I'd incorrectly had the RCodeselector set to "crc" not "md5"
Can you try opening the file with prolib.jar ? I'd like to know if that's a problem from the PL reader or from the rcode reader.
Otherwise, you can upload the two PL here: https://upload.rssw.eu, and I'll try to read the content (don't know *when* however...)
Will do, thanks :-)
OK, so I ran prolib.jar against my library and the MD5's look OK;
For eaxample RCODe file 450.r has the MD5 below..
CRC MD5 Size File
10106 6D73ACC29364096060F84B0670F514FD 91003 450.r
And when I run my own RCODE MD5 checker I get this (after extracting 450.r from the lib)...
---------------------------
Information
---------------------------
Attribute: 1st 2nd RCODE-INFO
CODEPAGE: ISO8859-1 ISO8859-1
CRC-VALUE: 10106 10106
DB-REFS: cdk cdk
FILE-NAME: \build\rcode\temp\450.r \build\library\450.r
HANDLE: 25 25
INST-PROC: ? ?
IS-CLASS: no no
LANGUAGES: <unnamed> <unnamed>
MD5: 6D73ACC29364096060F84B0670F514FD 6D73ACC29364096060F84B0670F514FD
TBL-CRCS: 45507,45507,4759 45507,45507,4759
TYPE: PSEUDO-WIDGET PSEUDO-WIDGET
MATCH: yes
---------------------------
OK
---------------------------
So the problem probably comes from the new rcode; would you mind sending one of the rcode just generated so that I can verify they can be read correctly ?
Hi,
Do you know if the latest PCT.jar has fixed this issue?
Thanks,
Mark.