Description of problem: The script runs needing user intervention -- this prevents running it unattended, as in a cron process to package new versions which appear on the mirrors This is the offending text: Do you want to move the source (Bchron_2.1.tar.gz) in /home/herrold/rpmbuild/SOURCES/? Y/N [Y] Version-Release number of selected component (if applicable): R2spec-2.5.1-6orc How reproducible: run the script and see that it needs interaction Steps to Reproduce: as above Actual results: as above Expected results: Adding an option '-y' to answer questions 'yes' as does 'yum' would avoid the need for interaction Additional info: Please ask if you have any questions
Seems good to me and its a easy fix Watch out that : -Y => If the spec is already present that will overwrite it => Will copy the sources into the appropriate folder What about a -c --copySource option instead ? ---------- BTW should it remove the ./source.tar.gz ? IE : Do you want to copy ? if Yes: Copy file toFolder rm file PS Would you mind put the RFE on the r2spec website directly I should receive the emails directly
I wrote a couple of workaround acripts .. shell and expect, to bypass the issue temporarily: [herrold@centos-5 R-packages]$ cat README ; \ echo "=======================" ; \ cat script.exp.template #!/bin/sh # # get all R project packages QUIET="-q " # for i in ` lynx -dump http://cran.r-project.org/web/packages/ | \ grep "packages" | awk {'print $2'} | grep ""index.html$` ; do for j in `lynx -dump $i | grep z$ | grep src | awk {'print $2'} `; do BAS=`basename $j` [ ! -e $BAS ] && { URL=`echo "$i "` wget $QUIET $j } [ -e $BAS -a "x${URL}" != "x" ] && { TGT=`echo "$BAS" | awk -F_ {'print $1'}` SRCS=`echo "$j "` # # clean up from any prior run, successful or not [ -e R-${TGT}.spec ] && rm -f R-${TGT}.spec [ -d tmp ] && rm -rf tmp # # R2spec --cran -s ${BAS} sed -e "s#@SOURCE@#${BAS}#g" < script.exp.template > ${TGT}.expect chmod 755 ${TGT}.expect ./${TGT}.expect } # # post run cleanup touch index.html rm -f index.html* done done # ======================= #!/usr/bin/expect -f # from autoexpect set force_conservative 0 ;# set to 1 to force conservative mode even if set timeout -1 spawn R2spec --cran -s @SOURCE@ match_max 100000 expect -exact "/SOURCES/? Y/N \[Y\]\r " send -- "\r" expect eof [herrold@centos-5 R-packages]$
https://fedorahosted.org/r2spec/browser/devel/CHANGELOG
as to options either way is fine ... sp long as it is documented, and permits scripting for hands off use, I am happy; the other places which shoulod be converted to options them becmoe the 'n (name) and -e (email) , and -a (archive) target (BIOC v CRAN) ... I package and build everything as non-root of course, and and dislike needing to ask an end user to have to contact a root level priv person --- some sites to not like such interaction ;( ==================== copy of mv ... either way ... I would tend to NOT have the default mv, as it may confuse other processes ... like the one I pasted a moment ago -- it would unnecessarily re-download a tarball it already had, if mv were a default Also and rpmbuild -ba (spec) will carry the tarball inside the resulting SRPM it makes; managing ~/rpmbuild/SOURCES/ and ~rpmbuild/SRPMS/ are out of scope for R2spec, as I see it. ==================== as to the PS as to RFE's only ... I would strongly prefer to have a central place for working on packages ... the Red Hat bugzilla has many hundreds of bugs reported by me over the years, and having to remember a exception for just a specific package is troublesome. If email addresses used for you by Bugzilla for packaged you maintain are a problem, why not alter the email address it uses to one you prefer? That way one change one place one time fixes everything. -- Russ herrold
(In reply to comment #4) > as to options > > either way is fine ... sp long as it is documented, and permits scripting for > hands off use, I am happy; the other places which shoulod be converted to > options them becmoe the 'n (name) and -e (email) , and -a (archive) target > (BIOC v CRAN) As you can see from the CHANGELOG I already have implemented the -c/--copyFile. It would indeed make sense to have also as option the email, the name and I would keep the --cran/--bioc because then I can use --bioc=Version > > ... I package and build everything as non-root of course, and and dislike > needing to ask an end user to have to contact a root level priv person --- some sites to not like such interaction ;( I am not sure what you mean here :/ (sorry about that) > > ==================== > > copy of mv ... either way ... I would tend to NOT have the default mv, as it > may confuse other processes ... like the one I pasted a moment ago -- it would > unnecessarily re-download a tarball it already had, if mv were a default > > Also and rpmbuild -ba (spec) will carry the tarball inside the resulting SRPM > it makes; managing ~/rpmbuild/SOURCES/ and ~rpmbuild/SRPMS/ are out of scope > for R2spec, as I see it. Indeed, then I'd rather let it copy the source rather than move them :) > > ==================== > > as to the PS as to RFE's only ... I would strongly prefer to have a central > place for working on packages ... the Red Hat bugzilla has many hundreds of > bugs reported by me over the years, and having to remember a exception for just > a specific package is troublesome. > > If email addresses used for you by Bugzilla for packaged you maintain are a > problem, why not alter the email address it uses to one you prefer? That way > one change one place one time fixes everything. I was not an issue of the email address but more a way for me to also keep things together, but it does not really matter Let's stay here I am also fine with that :)
Options -n and -e implemented https://fedorahosted.org/r2spec/browser/devel/CHANGELOG
> It was not an issue of the email address but more a way for me to also keep > things together, but it does not really matter > Let's stay here I am also fine with that :) Thank you
This bug has been triaged