Hi Klaus, Below is a little test I put together to try to help you out: avp_write("<tel:+43108>", "$avp(s:pai)"); avp_subst("$avp(s:pai)/$avp(s:cli)","/<tel\:(.*)>/\1/"); xlog("L_NOTICE", "pai=$avp(s:pai) cli=$avp(s:cli)\n"); This is the result: pai=<tel:+43108> cli=+43108

Regards, Norm Klaus Darilion wrote:

Hi Klaus, regexp tries to do the longest match for each token so the (.*) will include also the trailing spaces. try not to put ., but everything except spaces, like [^ ]* regards, bogdan Klaus Darilion wrote:

no other clue in this case :( regards, bogdan Klaus Darilion wrote:

Hi Norman! thanks - this works indeed. Now I wanted to tune it, so that it not only removes spaces, but also tabs (according to posix this is [:blank:]). For positive matching it works: /<[[:blank:]]*tel:([^ ]+)[[:blank:]]*>/\1/ --> OK But if I add this to the "negative matching" it fails: /<[[:blank:]]*tel:([^[[:blank:]]]+)[[:blank:]]*>/\1/ --> not OK ^^^^^^^^^^^^ Do I need a special syntax for this? thanks klaus btw: I think escaping the colon is not necessary Norman Brandinger wrote:

Hi! Meanwhile I splitted the regexp into several smaller ones nad now it works. But I found an interesting thing when removing blanks from an AVP: This one works fine: avp_subst("$avp(s:pai)/$avp(s:painb)","/[[:blank:]]//g"); This one causes "out of mem": avp_subst("$avp(s:pai)/$avp(s:painb)","/[[:blank:]]*//g"); ^ 07:17:46 server2 /usr/sbin/openser[6565]: subst_run: running. r=0 07:17:46 server2 /usr/sbin/openser[6565]: subst_run: matched (0, 0): [] 07:18:12 server2 /usr/sbin/openser[6565]: subst_run: running. r=0 07:18:12 server2 /usr/sbin/openser[6565]: subst_run: matched (0, 0): [] 07:18:12 server2 /usr/sbin/openser[6565]: subst_run: running. r=0 07:18:12 server2 /usr/sbin/openser[6565]: subst_run: matched (0, 0): [] 07:18:12 server2 /usr/sbin/openser[6565]: ERROR: replace_build: out of mem (rpl) 07:18:12 server2 /usr/sbin/openser[6565]: subst_str: no match 07:18:12 server2 /usr/sbin/openser[6565]: avpops:ops_subst: subst to 0 avps regards klaus Klaus Darilion wrote: > Norman Brandinger wrote: > >> Hi Klaus, >> >> Below is a little test I put together to try to help you out: >> >> avp_write("<tel:+43108>", "$avp(s:pai)"); >> avp_subst("$avp(s:pai)/$avp(s:cli)","/<tel\:(.*)>/\1/"); >> xlog("L_NOTICE", "pai=$avp(s:pai) cli=$avp(s:cli)\n"); >> >> This is the result: >> >> pai=<tel:+43108> cli=+43108 > > > Hi Norman. Yes, you are right. I simply had a wrong regular > expression. But now I have another problem: > > Input: $avp(s:pai) = '< tel:+43103 >' > (the part within '') > > avp_subst("$avp(s:pai)/$avp(s:paias)","/.*<[[:blank:]]*(.*?)[[:blank:]]*>.*/\1/") > > > Output: $avp(s:paias) = 'tel:+43103 ' > > Thus, the leading whitespace is removed, but not the trailing one, > but I can't find an error in my regular expression. Maybe some regexp > experts out there? > > thanks > klaus > >> >> Regards, >> Norm >> >> Klaus Darilion wrote: >> >>> Hi! >>> >>> I have problems with avp_subst: >>> >>> I have $avp(s:pai) with the value <tel:+43108> (the < and > belongs >>> to the AVP) >>> >>> Then I want to extract the number into $avp(s:cli): >>> >>> avp_subst("$avp(s:pai)/$avp(s:cli)","/tel:(.*)/\1/") >>> >>> The result is <+43108> >>> >>> Can someone explain me why the leading < belongs to the result? >>> Shouldn't it be removed? >>> >>> thanks >>> klaus >>