Bump this thread when you fix a bug

diff -r d080bed0babc mpf1t3xy.pro
--- a/mpf1t3xy.pro      Tue Jun 09 23:24:46 2009 +0100
+++ b/mpf1t3xy.pro      Tue Jun 09 23:25:34 2009 +0100
@@ -19,7 +19,7 @@
     intercept = p[1]
     f = slope * x + intercept
     if n_elements(e_int) eq 0 then e_int = 0.0
-    resid = (y - f)/(e_y^2 + slope^2*e_x^2 + e_int^2)
+    resid = (y - f)/sqrt((e_y^2 + slope^2*e_x^2 + e_int^2))
     return, resid
 end

CHECK THAT IN

This had been giving me jokes for two months. Had been testing mpf1t3xy2 function all that time. Finally noticed that I was calling the function wrongly at 2am last night.

bembo:tully$ hg diff -r 608 tfall.pro 
diff -r 2bfbd2cb20a2 tfall.pro
--- a/tfall.pro Wed Jun 10 00:41:50 2009 +0100
+++ b/tfall.pro Thu Jun 11 15:12:22 2009 +0100
@@ -7,15 +7,6 @@
     sp = ['NGC_3390', 'PGC_44931', 'ESO_443G42', 'NGC_5746', 'IC_4767', $
         'NGC_6722', 'ESO_185G53', 'IC_4937', 'IC_5096', 'ESO_240G11', $
         'NGC_1886', 'NGC_4703', 'NGC_7123', 'IC_5176']
-
-    ; Old galaxy order (FIX: this gives different answers for log mass fits,
-    ; and possibly others. This may be a bug in mpf1t3xy)
-    ;s0 = ['NGC_0128', 'NGC_1381', 'NGC_1596', 'NGC_2310', 'ESO_311G12', $
-    ;        'NGC_3203', 'NGC_4469', 'NGC_4710', 'NGC_6771', 'ESO_597G36', $
-    ;        'NGC_1032', 'NGC_3957', 'NGC_5084', 'ESO_151G04']
-    ;sp = ['ESO_185G53', 'ESO_240G11', 'ESO_443G42', 'IC_4767', 'IC_4937', $
-    ;        'IC_5096', 'IC_5176', 'NGC_1886', 'NGC_3390', 'NGC_4703', $
-    ;        'NGC_5746', 'NGC_6722', 'NGC_7123', 'PGC_44931']
 
     ; Fix distance and apparent magnitude errors
     fixderror = 0.0
@@ -113,32 +104,32 @@
 
     print, "log v_obs -- log mass (stellar)"
     logv_logmass = mpf1t3xy2(logvsp, logmasssp, logvs0, logmasss0, $
-        e_logvsp, e_logmasssp, e_logvs0, e_logmasssp, /quiet, x0 = x0, $
+        e_logvsp, e_logmasssp, e_logvs0, e_logmasss0, /quiet, x0 = x0, $
         guess = vmassguess, reduce = reduce, /latex)
 
     print, "log v_drift -- log mass (stellar)"
     logvdrift_logmass = mpf1t3xy2(logvdriftsp, logmasssp, logvdrifts0, $
-        logmasss0, e_logvdriftsp, e_logmasssp, e_logvdrifts0, e_logmasssp, $
+        logmasss0, e_logvdriftsp, e_logmasssp, e_logvdrifts0, e_logmasss0, $
         /quiet, x0 = x0, guess = vmassguess, reduce = reduce, /latex)
 
     print, "log v_c -- log mass (stellar)"
     logvc_logmass = mpf1t3xy2(logvcsp, logmasssp, logvcs0, logmasss0, $
-        e_logvcsp, e_logmasssp, e_logvcs0, e_logmasssp, /quiet, x0 = x0, $
+        e_logvcsp, e_logmasssp, e_logvcs0, e_logmasss0, /quiet, x0 = x0, $
         guess = vmassguess, reduce = reduce, /latex)
 
     print, "log v_obs -- log mass (dynamical)"
     logv_logmassdyn = mpf1t3xy2(logvsp, logmassspdyn, logvs0, logmasss0dyn, $
-        e_logvsp, e_logmassspdyn, e_logvs0, e_logmassspdyn, /quiet, x0 = x0, $
+        e_logvsp, e_logmassspdyn, e_logvs0, e_logmasss0dyn, /quiet, x0 = x0, $
         guess = vmassguess, reduce = reduce, /latex)
 
     print, "log v_drift -- log mass (dynamical)"
     logvdrift_logmassdyn = mpf1t3xy2(logvdriftsp, logmassspdyn, logvdrifts0, $
         logmasss0dyn, e_logvdriftsp, e_logmassspdyn, e_logvdrifts0, $
-        e_logmassspdyn, /quiet, x0 = x0, guess = vmassguess, reduce = reduce, /latex)
+        e_logmasss0dyn, /quiet, x0 = x0, guess = vmassguess, reduce = reduce, /latex)
 
     print, "log v_c -- log mass (dynamical)"
     logvc_logmassdyn = mpf1t3xy2(logvcsp, logmassspdyn, logvcs0, logmasss0dyn, $
-        e_logvcsp, e_logmassspdyn, e_logvcs0, e_logmassspdyn, /quiet, x0 = x0, $
+        e_logvcsp, e_logmassspdyn, e_logvcs0, e_logmasss0dyn, /quiet, x0 = x0, $
         guess = vmassguess, reduce = reduce, /latex)

I fixed loads of things in ILX

diffs or it's not true

lol not being using subversion for a while so no diffs, so it's not true.

im a tester so i find bugs but no fixies

You mean with my stuff, or your work?

wait is there ANY kind of version control with ilx????

work - only do it for $$$$

Well, I did version control it with svn for a while, but given that almost all of it is written just by me, there's precious little point. Having backups is important though.

(x-post) yes, can't expect fixes quickly at work.

ilx version control

http://memory.loc.gov/ammem/collections/whitman/images/notebook2.jpg

In that, the only reason I would want version control is to maybe roll back an ill-advised change I had made to it... Which is something that basically never happens, but that I could do by using various backup mechanisms.

I don't fix bugs. I create workarounds.

On a related note, C++Builder 2009 needs to fucking die.

im a tester so i find bugs but no fixies

lol me 2

Ohh man array.sort gives different results depending on whether the contents of the array have themselves been sorted!

YOU WERE A WORTHY ADVERSARY, BUG*

(still not sure what caused them to be in different order in the first place but whatev)

im off work until jan 18 - NO BUGS FOR ME <3<3<3<3

i'm drowning in bugs due to developer incompetence ;_;

bugs are gone! some of my logic was actually correct the first time!

Holy shit there is some funky HTML in old-old-ILX comments. More than 188 posts with <HTML> in them, which threw up all sorts of parser bugs for me.

'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
'
' ???? !!!!
' ( *_*) ERROR HANDLING (^_^ )
'
'
'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

bah

i think there's a bug in your post

diff -r 63cbed0e502c plotrmass.pro
--- a/plotrmass.pro Wed Nov 04 18:46:34 2009 +0000
+++ b/plotrmass.pro Wed Nov 04 22:22:23 2009 +0000
@@ -26,12 +26,15 @@
logmasss0dyn = logmass(s0, error = e_logmasss0dyn, fixapperror = fixapperror, $
fixderror = fixderror, /dyn)

- rds0 = alog10(getrd(s0))
- rdsp = alog10(getrd(sp))
- res0 = alog10(getre(s0))
- resp = alog10(getre(sp))
- r25s0 = alog10(getr25(s0))
- r25sp = alog10(getr25(sp))
+ ; Must be in physical units not arcsec!
+ ; pc = distance*!dpi/(648d) ; Constant factor to convert arcsec --> kpc
+
+ rds0 = alog10(getrd(s0) * getdistance(s0)*!dpi/648d)
+ rdsp = alog10(getrd(sp) * getdistance(sp)*!dpi/648d)
+ res0 = alog10(getre(s0) * getdistance(s0)*!dpi/648d)
+ resp = alog10(getre(sp) * getdistance(sp)*!dpi/648d)
+ r25s0 = alog10(getr25(s0) * getdistance(s0)*!dpi/648d)
+ r25sp = alog10(getr25(sp) * getdistance(sp)*!dpi/648d)

IT'S ALWAYS DISTANCE

FUCKING ANGULAR SIZE

hey didn't you make the Kessel Run in 12 parsecs arcseconds?

people who measure the size of objects at known distance in angular size are the most disgusting savages in the world imo

this FF bug title made me laff for some reason

https://bugzilla.mozilla.org/show_bug.cgi?id=468562

i can't do a diff for this one, but fucking white dwarf spectra.

it just took me ~6 hours to figure out there was nothing wrong in my onclick checkbox javascript function... my text label div was covering it so i couldn't click. feel like cheering/crying/puking

9136 bembo:idl$ hg diff ppxf_bootstrap.pro 
diff -r d74b7b28df85 ppxf_bootstrap.pro
--- a/ppxf_bootstrap.pro        Tue Mar 02 16:59:17 2010 +0100
+++ b/ppxf_bootstrap.pro        Tue Mar 02 23:17:13 2010 +0100
@@ -174,15 +174,15 @@
             goodpixels = goodpixels1, moments = 4, degree = degree, $
             vsyst = vsyst, bias = 0.5, bestfit = bestfit, plot = plot, $
             ps = ps, specname = frame + '_bin_' + strtrim(j, 2), error = error
-        resid = galaxy - bestfit
+        resid = galaxy_bin - bestfit

fucking variable names

if lte IE 6 do this
if !IE 6 do this instead

how does "if not IE 6" not apply to IE 8? very retarded imo.

got jmagick compiling, finally - the makefile was trying to remove base path from current working directory to give relative path but was failing as the directory was a softlink and the two different bits of code were returning different names for the same place. solution was to move everything under /tmp and compile it there, install the target files and just not worry about the src files.

(jmagick is 3 times faster than the current method of doing the same thing (resizing images dynamically). target is 100 a second in a clustered environment of hefty servers. i am currently getting 50 in .53 seconds on my laptop using only one core...)

if lte IE 6 do this
if !IE 6 do this instead

how does "if not IE 6" not apply to IE 8? very retarded imo.

― bnw, Friday, April 9, 2010 12:48 PM

iirc there is no "!IE" conditional statement, so use "if gt IE 6"

if ur talking css

if IE gt fo

    @Path("/track={purchaseId:{[0-9a-f]{32}}")

    @Path("/track={purchaseId:[0-9a-f]{32}}")

one works, the other fails silently. bloody annotations. bloody frameworks.

fixed a bug in an xsl transform, using a workaround i found in a list post from 2003, that even then was described as only being necessary for old xsl transformation engines.

a daily wtf:

convert string to byte array, pass byte array to save method
in save method, convert byte array back to string, write string to file from 0 to bytearray.length

maybe useful if the save method takes a length in bytes. however it takes a length in characters. these are not necessarily the same. hello indexoutofbounds exception, at least a week of hair-pulling over assuming it was a problem with dodgy non utf-8 characters.

was going to post some Excel VBA mysteries/irritations but really Excel VBA is one big bug in itself

nevertheless, it is the only thing I can do that nobody else in the office has ever wanted to dirty their brain by looking into

(most annoying this week: using Application.Workbooks.OpenText to open a text file, specifying all the parameters, except if the file extension is .csv Excel overrides them without telling you. So I'm going to have to write some documentation that says "look, I know the file is a .csv file, but please call it .txt or anything except .csv or it will load wrongly in a way that you may not notice until it's too late")

how long will it take to learn that the problem may be in the data and not the programming... 2 hours of trying to locate an extra space in a string..............urgh

strange things on the interweb:

SELECT ThePerfectWoman
FROM AllTheWomenInTheWorld
WHERE HerInterests Like Mine
AND CurrentMate Is Null
AND CurrentlyLives IN ('Yonkers', 'Bondi Beach', 'São Paulo')
GROUP BY CASE WHEN AsManyAsWillHaveMe>AsManyAsMyStaminaWillAllow
THEN AsManyAsWillHaveMe ELSE AsManyAsMyStaminaWillAllow * 2 END
HAVING Max(Libido)>=Mine

lol, yonkers