Jonathan H
2016-Oct-17 13:31 UTC
[asterisk-users] Multiple readfile oddities, newlines etc
I have a plain text file, ASCII, unix line breaks. 1 single line, and all that is in it is the word "radio". Here's some test dialplan: exten => 5,1,Verbose(Context: ${CONTEXT} Exten:${EXTEN}) same => n,Set(feature=${FILE(/home/test/feature-1.txt,0,1,l,u)}) same => n,Verbose(${feature}) same => n,Set(featurefile=/home/test/feature-1.txt) same => n,Set(feature2=${FILE(${featurefile},0,1,l,u)}) same => n,Verbose(${feature2}) Both should output "radio", right? Here's the output: -- Executing [5 at fromvoipfone201:2] Set("PJSIP/6001-00000052", "feature=radio") in new stack -- Executing [5 at fromvoipfone201:3] Verbose("PJSIP/6001-00000052", "radio") in new stack radio -- Executing [5 at fromvoipfone201:4] Set("PJSIP/6001-00000052", "featurefile=/home/test/feature-1.txt") in new stack -- Executing [5 at fromvoipfone201:5] Set("PJSIP/6001-00000052", "feature2*radi*") in new stack *GAAAAARRRGGGHHHH! ^^^^^* -- Executing [5 at fromvoipfone201:6] Verbose("PJSIP/6001-00000052", "radi") in new stack And this is what's just at the top of the script. If I put it way down in another context: -- Executing [s at track-handler:3] Verbose("Local/s at root-00000026;2", "/home/test/feature-1") in new stack /home/test/feature-1 -- Executing [s at track-handler:4] Set("Local/s at root-00000026;2", "feature=*radio??*") in new stack *EVEN MORE** GAAAAARRRGGGHHHH! ^^^^^* [Oct 17 13:29:33] ERROR[5093][C-0000003c]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 13:29:33] ERROR[5093][C-0000003c]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message -- Executing [s at track-handler:5] Verbose("Local/s at root-00000026;2", "radio??") in new stack radio?? -- Executing [s at track-handler:6] GotoIf("Local/s at root-00000026;2", "1?radio??,s,1") in new stack So, at the top of the file, when it's just a straight file as the filename, it gives the full word. If the filename is a variable, it strips and character. And further down, same thing, but it adds two weird blocks, which appear side by side in the console, but weirdly, stacked up in this email. (Oh, by the way, if there is more than one line in the file, even if I used "u" for line breaks, it adds a newline to the variable, Is that correct?) What am I missing? I've opened it in both nano and notepad++, I've used iconv and all the tools I can think of to check that file, and all the asterisk conf files, too. It all looks as it should here: $ cat feature-1.txt radio$ file feature-1.txt feature-1.txt: ASCII text $ wc -l feature-1.txt 0 feature-1.txt wc -c feature-1.txt 5 feature-1.txt After 6 hours struggling with this, I think I'm starting to lose the plot. Can anyone tell me where I'm going wrong? Thanks. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20161017/392f13bd/attachment.html>
Sebastian Nielsen
2016-Oct-17 13:50 UTC
[asterisk-users] Multiple readfile oddities, newlines etc
Theres always garbage in the end of the files. I do this when I want to read a file: same => n,Set(featurefile=/home/test/feature-1.txt) same => n,Set(unfilteredfeat2=${FILE(${featurefile},0,1,l,u)}) same => n,Set(feature2=${SHIFT(unfilteredfeat2)}) After that, add a , inside end of the file, so Cat feature-1.txt Reads: radio, Thus if there is garbage in the file, it will happen after the ,. Same if garbage happens to find its way into the end of the variable for some reason. Fr?n: asterisk-users-bounces at lists.digium.com [mailto:asterisk-users-bounces at lists.digium.com] F?r Jonathan H Skickat: den 17 oktober 2016 15:32 Till: Asterisk Users Mailing List - Non-Commercial Discussion <asterisk-users at lists.digium.com> ?mne: [asterisk-users] Multiple readfile oddities, newlines etc I have a plain text file, ASCII, unix line breaks. 1 single line, and all that is in it is the word "radio". Here's some test dialplan: exten => 5,1,Verbose(Context: ${CONTEXT} Exten:${EXTEN}) same => n,Set(feature=${FILE(/home/test/feature-1.txt,0,1,l,u)}) same => n,Verbose(${feature}) same => n,Set(featurefile=/home/test/feature-1.txt) same => n,Set(feature2=${FILE(${featurefile},0,1,l,u)}) same => n,Verbose(${feature2}) Both should output "radio", right? Here's the output: -- Executing [5 at fromvoipfone201:2] Set("PJSIP/6001-00000052", "feature=radio") in new stack -- Executing [5 at fromvoipfone201:3] Verbose("PJSIP/6001-00000052", "radio") in new stack radio -- Executing [5 at fromvoipfone201:4] Set("PJSIP/6001-00000052", "featurefile=/home/test/feature-1.txt") in new stack -- Executing [5 at fromvoipfone201:5] Set("PJSIP/6001-00000052", "feature2=radi") in new stack GAAAAARRRGGGHHHH! ^^^^^ -- Executing [5 at fromvoipfone201:6] Verbose("PJSIP/6001-00000052", "radi") in new stack And this is what's just at the top of the script. If I put it way down in another context: -- Executing [s at track-handler:3] Verbose("Local/s at root-00000026;2", "/home/test/feature-1") in new stack /home/test/feature-1 -- Executing [s at track-handler:4] Set("Local/s at root-00000026;2", "feature=radio??") in new stack EVEN MORE GAAAAARRRGGGHHHH! ^^^^^ [Oct 17 13:29:33] ERROR[5093][C-0000003c]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 13:29:33] ERROR[5093][C-0000003c]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message -- Executing [s at track-handler:5] Verbose("Local/s at root-00000026;2", "radio??") in new stack radio?? -- Executing [s at track-handler:6] GotoIf("Local/s at root-00000026;2", "1?radio??,s,1") in new stack So, at the top of the file, when it's just a straight file as the filename, it gives the full word. If the filename is a variable, it strips and character. And further down, same thing, but it adds two weird blocks, which appear side by side in the console, but weirdly, stacked up in this email. (Oh, by the way, if there is more than one line in the file, even if I used "u" for line breaks, it adds a newline to the variable, Is that correct?) What am I missing? I've opened it in both nano and notepad++, I've used iconv and all the tools I can think of to check that file, and all the asterisk conf files, too. It all looks as it should here: $ cat feature-1.txt radio$ file feature-1.txt feature-1.txt: ASCII text $ wc -l feature-1.txt 0 feature-1.txt wc -c feature-1.txt 5 feature-1.txt After 6 hours struggling with this, I think I'm starting to lose the plot. Can anyone tell me where I'm going wrong? Thanks. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20161017/55fd86ad/attachment.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 6298 bytes Desc: S/MIME Cryptographic Signature URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20161017/55fd86ad/attachment.bin>
Jonathan H
2016-Oct-17 14:49 UTC
[asterisk-users] Multiple readfile oddities, newlines etc
Just by chance, I was browsing the mailing list and it looks like Sebastian's reply made it to the list, but not to me because it's apparently blacklisted by UCEPROTECTL3. This is what he wrote:> Theres always garbage in the end of the files. > I do this when I want to read a file: > same => n,Set(featurefile=/home/test/feature-1.txt) > same => n,Set(unfilteredfeat2=${FILE(${featurefile},0,1,l,u)}) > same => n,Set(feature2=${SHIFT(unfilteredfeat2)}) > After that, add a , inside end of the file, so > Cat feature-1.txt > Reads: > radio, > Thus if there is garbage in the file, it will happen after the ,. Same if garbage happens to find its way into the end of the variable for some reason.Well, OK, that sort of works, but it revealed even more odd things. And, if tools like wc are saying that a file is a certain length, how can there be junk hiding in it? So, then I found that having different lengths in the file messed things up, too. Can anyone tell me this is normal?!? Have a look, this is for changing the following words in the features.txt. In all of the examples, the LEN function gets it wrong, too. radio < gets cut a letter short even with SHIFT trimmer radio, < works, but only by chance because , makes it 6 characters radui12345 < garbage at end radio12345, < shift trimmer removes garbage rad < junk at end Context: fromvoipfone201 Exten:5 Using a string, feature is set to ---radio--- and is 7 long Using a variable, feature2 is set to ---radi--- and is 8 long Using a string with shift method, feature3 is set to ---radi--- and is 8 long Context: fromvoipfone201 Exten:5 Using a string, feature is set to ---radio,--- and is 7 long Using a variable, feature2 is set to ---radio,--- and is 8 long Using a string with shift method, feature3 is set to ---radio--- and is 8 long Context: fromvoipfone201 Exten:5 Using a string, feature is set to ---radio12345--- and is 7 long [Oct 17 15:30:29] ERROR[5813][C-00000073]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:29] ERROR[5813][C-00000073]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message Using a variable, feature2 is set to ---radio12345G?--- and is 8 long [Oct 17 15:30:29] ERROR[5813][C-00000073]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:29] ERROR[5813][C-00000073]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message [Oct 17 15:30:29] ERROR[5813][C-00000073]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:29] ERROR[5813][C-00000073]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message Using a string with shift method, feature3 is set to ---radio12345G?--- and is 8 long Context: fromvoipfone201 Exten:5 Using a string, feature is set to ---radio12345,--- and is 7 long [Oct 17 15:30:35] ERROR[5814][C-00000074]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:35] ERROR[5814][C-00000074]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message Using a variable, feature2 is set to ---radio12345,?--- and is 8 long [Oct 17 15:30:35] ERROR[5814][C-00000074]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:35] ERROR[5814][C-00000074]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message [Oct 17 15:30:35] ERROR[5814][C-00000074]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:35] ERROR[5814][C-00000074]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message Using a string with shift method, feature3 is set to ---radio12345--- and is 8 long Context: fromvoipfone201 Exten:5 Using a string, feature is set to ---rad--- and is 7 long [Oct 17 15:30:43] ERROR[5815][C-00000075]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:43] ERROR[5815][C-00000075]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message Using a variable, feature2 is set to ---rad?--- and is 8 long [Oct 17 15:30:43] ERROR[5815][C-00000075]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:43] ERROR[5815][C-00000075]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message [Oct 17 15:30:43] ERROR[5815][C-00000075]: json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string. [Oct 17 15:30:43] ERROR[5815][C-00000075]: stasis_channels.c:773 ast_channel_publish_varset: Error creating message Using a string with shift method, feature3 is set to ---rad?--- and is 8 long So, using the SHIFT trimmer cuts the junk, but still loses the last character in some cases. Here's the dialplan same => n,Set(feature=${FILE(/home/test/feature-1.txt,0,1,l,u)}) same => n,Verbose(1,Using a string, feature is set to ---${feature}--- and is ${LEN(feature)} long) same => n,Set(featurefile=/home/test/feature-1.txt) same => n,Set(feature2=${FILE(${featurefile},0,1,l,u)}) same => n,Verbose(1,Using a variable, feature2 is set to ---${feature2}--- and is ${LEN(feature2)} long) same => n,Set(unfilteredfeat=${FILE(${featurefile},0,1,l,u)}) same => n,Set(feature3=${SHIFT(unfilteredfeat)}) same => n,Verbose(1,Using a string with shift method, feature3 is set to ---${feature3}--- and is ${LEN(feature3)} long) Bug or... "feature"?! On 17 October 2016 at 14:31, Jonathan H <lardconcepts at gmail.com> wrote:> I have a plain text file, ASCII, unix line breaks. 1 single line, and all > that is in it is the word "radio". > > Here's some test dialplan: > > exten => 5,1,Verbose(Context: ${CONTEXT} Exten:${EXTEN}) > > same => n,Set(feature=${FILE(/home/test/feature-1.txt,0,1,l,u)}) > same => n,Verbose(${feature}) > > same => n,Set(featurefile=/home/test/feature-1.txt) > same => n,Set(feature2=${FILE(${featurefile},0,1,l,u)}) > same => n,Verbose(${feature2}) > > Both should output "radio", right? Here's the output: > > -- Executing [5 at fromvoipfone201:2] Set("PJSIP/6001-00000052", > "feature=radio") in new stack > -- Executing [5 at fromvoipfone201:3] Verbose("PJSIP/6001-00000052", > "radio") in new stack > radio > > -- Executing [5 at fromvoipfone201:4] Set("PJSIP/6001-00000052", > "featurefile=/home/test/feature-1.txt") in new stack > -- Executing [5 at fromvoipfone201:5] Set("PJSIP/6001-00000052", > "feature2=radi") in new stack > GAAAAARRRGGGHHHH! > ^^^^^ > > -- Executing [5 at fromvoipfone201:6] Verbose("PJSIP/6001-00000052", > "radi") in new stack > > And this is what's just at the top of the script. If I put it way down in > another context: > > -- Executing [s at track-handler:3] Verbose("Local/s at root-00000026;2", > "/home/test/feature-1") in new stack > /home/test/feature-1 > -- Executing [s at track-handler:4] Set("Local/s at root-00000026;2", > "feature=radio??") in new stack > EVEN MORE > GAAAAARRRGGGHHHH! ^^^^^ > > [Oct 17 13:29:33] ERROR[5093][C-0000003c]: json.c:704 ast_json_vpack: Error > building JSON from '{s: s, s: s}': Invalid UTF-8 string. > [Oct 17 13:29:33] ERROR[5093][C-0000003c]: stasis_channels.c:773 > ast_channel_publish_varset: Error creating message > -- Executing [s at track-handler:5] Verbose("Local/s at root-00000026;2", > "radio??") in new stack > radio?? > -- Executing [s at track-handler:6] GotoIf("Local/s at root-00000026;2", > "1?radio??,s,1") in new stack > > > So, at the top of the file, when it's just a straight file as the filename, > it gives the full word. If the filename is a variable, it strips and > character. And further down, same thing, but it adds two weird blocks, which > appear side by side in the console, but weirdly, stacked up in this email. > > (Oh, by the way, if there is more than one line in the file, even if I used > "u" for line breaks, it adds a newline to the variable, Is that correct?) > > What am I missing? I've opened it in both nano and notepad++, I've used > iconv and all the tools I can think of to check that file, and all the > asterisk conf files, too. > It all looks as it should here: > > $ cat feature-1.txt > radio$ file feature-1.txt > feature-1.txt: ASCII text > $ wc -l feature-1.txt > 0 feature-1.txt > wc -c feature-1.txt > 5 feature-1.txt > > After 6 hours struggling with this, I think I'm starting to lose the plot. > Can anyone tell me where I'm going wrong? Thanks.
Pete Mundy
2016-Oct-17 21:10 UTC
[asterisk-users] Multiple readfile oddities, newlines etc
> On 18/10/2016, at 2:31 am, Jonathan H <lardconcepts at gmail.com> wrote: > > I have a plain text file, ASCII, unix line breaks. 1 single line, and all that is in it is the word "radio". > > <snip>Heya Jonathan Interesting problem! Unfortunately I can't help with suitable dialplan code to resolve your issue, but I thought I'd at least throw this snippet out to you: If you want to know what is _really_ in that file (including all invisible characters and anything else that wc etc might not count), pipe it through 'hexdump'. Then you get to see the raw contents of the file in hex and can compare it to other files or the output you get from other tools (and you can look up the ASCII char codes for invisible chars to explore what they are etc). Eg: cat /home/test/feature-1.txt | hexdump Hope this tool helps you in your quest :) Pete -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20161018/ca9d96ed/attachment.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3577 bytes Desc: not available URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20161018/ca9d96ed/attachment.bin>
Steve Edwards
2016-Oct-17 21:38 UTC
[asterisk-users] Multiple readfile oddities, newlines etc
On Tue, 18 Oct 2016, Pete Mundy wrote:> If you want to know what is _really_ in that file (including all > invisible characters and anything else that wc etc might not count), > pipe it through 'hexdump'. > > cat?/home/test/feature-1.txt | hexdumpOr just: hexdump /home/test/feature-1.txt -- Thanks in advance, ------------------------------------------------------------------------- Steve Edwards sedwards at sedwards.com Voice: +1-760-468-3867 PST https://www.linkedin.com/in/steve-edwards-4244281
Possibly Parallel Threads
- Multiple readfile oddities, newlines etc
- json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string.
- json.c:704 ast_json_vpack: Error building JSON from '{s: s, s: s}': Invalid UTF-8 string.
- help with crash
- ARI strange bug on version 13.29.2